Copyright

Is the Copyright Act Inconsistent with the Law of Employee Invention Assignment Contracts?

Is the Copyright Act Inconsistent with the Law of Employee  Invention Assignment Contracts?

Charles Tait Graves*

Download a PDF version of this article here

There is a latent conflict between the law of employee invention assignment contracts and the Copyright Act’s work for hire doctrine. Countless employees sign contracts specifying that, in most cases, the employer will own trade secrets and patentable inventions, as well as copyrightable works. When employees create in the workplace, these rules are largely uncontroversial. But when employees create something outside the workplace for a new venture, there can be a conflict between these two areas of intellectual property law. The work for hire doctrine is more favorable to employee-ownership than the law of invention assignment contracts. As a perhaps surprising result, where an employee’s outside-the-workplace creation might constitute both a trade secret and a copyrightable work, these two ownership tests can point in opposite directions. Further, when an employee prevails as to copyright ownership, there are good reasons why that result precludes an employer’s conflicting claim to trade secret ownership in the same work. This friction on the boundaries of two areas of intellectual property law has important policy ramifications for employees who create intellectual property on the side, while planning for their next job. **

 

Introduction

This article considers the conflict between two commonplace but little-discussed areas of intellectual property law: the law of invention assignment contracts and the law of works for hire under the Copyright Act.

Both can encompass the same subject matter. Both offer tests to allocate ownership between an employer and an employee when the employee creates something. However, sometimes these tests are in tension, if not outright conflict. This conflict is particularly apparent in the context of employee mobility disputes; i.e. those that arise when someone creates intellectual property on the side, outside the workplace, before quitting and leaving for his or her next opportunity.

Employees planning a new venture face a tangle of potential tort, contract, and statutory claims brought by employers seeking to halt competition or to claim ownership of the departing employee’s ideas. Two such claims center on whether the former employer can claim a property right in ideas or work product that a departing employee created offsite, on his or her own time. This essay is the first to compare these related areas of law, tease out the potential conflicts in their tests for employee ownership, and identify the reasons one offers better protection than the other for departing employees.

The law of invention assignment contracts and the work for hire law under the Copyright Act provide the basis for two claims employers can bring to seek ownership of intellectual property an employee has created, whether at the workplace or outside the workplace. Both areas of law start with contract terms that purport to govern whether an employer or an employee owns intellectual property that an employee creates. One applies to trade secrets and patentable inventions. The other applies to copyrightable works of expression. The former, the law of invention assignment agreements, is largely a matter of state contract and state statutory law. The latter, the work for hire doctrine, is largely a matter of federal copyright law. Through these extra-contractual sources, both provide rules that allocate ownership of employee work product, including creative work done entirely apart from the employer’s business. Thus, while many may believe that ownership of employee-created work product is a straightforward matter of contract, that is not always the case when we focus on creative materials an employee has prepared outside the company, for purposes outside the company.

There is a marked contrast between these dueling areas of law. Invention assignment law is highly one-sided, favoring the employer, while work for hire case law is more balanced, and employees often prevail in establishing ownership of their copyrighted creations. This difference arises because the Copyright Act focuses on what an employee was hired to do and what he or she intended by creating the work at issue. The law of invention assignment contracts, by contrast, treats the employee’s role and mental state as irrelevant and instead asks whether the created work is related in some manner to the employer’s business – a broad definition the boundaries of which remain unclear.

Increasing this tension, the subject matter of these doctrines can overlap, which means that the law can dictate different ownership outcomes for the same employee work product. For example, whenever an employee creates a work of expression that is simultaneously a potential trade secret (such as software code), the law of invention assignment contracts and the law of works for hire are both in play. Ownership results may differ depending on the set of tests used to adjudicate a particular dispute. This article will analyze cases where the outcome might have differed if the court and litigants had analyzed a different body of law.

Where there is a conflict, copyright ownership appears to override a conflicting ownership claim to trade secret rights in the same work product. Specifically, where an employee prevails and a copyrightable work is not the employer’s work for hire, the employer likely cannot win by asserting trade secret rights. Because a copyright owner can disclose and publish the work at will, trade secret rights – which require secrecy – would be precluded from the outset. Thus, from a departing employee’s perspective, establishing copyright ownership may be paramount. That may well be better public policy given the potential for overreach in the law of invention assignment agreements.

To parse these areas of law which could govern ownership rights in the same subject matter, this article will explore both areas of law in depth, with a survey of virtually all of the available case law on ownership disputes between employers and employees. Providing this survey may help address issues that largely remain unexplored by courts and commentators, even those who are sympathetic to the obstacles departing employees can face.[1]

Section I explains that when employees create outside-the-workplace material for a future new venture, two categories of intellectual property law are relevant: the law of invention assignment contracts and work for hire law under the Copyright Act. This article will focus on employees acting in good faith to create for a potential new company, in order to set the stage for the latent clash between these doctrines.

Section II provides a historical overview of the development of invention assignment law. This runs from early common law rules and their important exceptions, to the modern era, where the combination of state invention assignment statutes and commonplace contract terms have substantially modified the common law. Setting up the conflict between the law of invention assignment contracts and work for hire law, this section also demonstrates how invention assignment case law significantly favors employers over employees.

Section III sets forth a contrasting body of law which also addresses employees’ creation of intellectual property, with sometimes different outcomes for disputes between employers and employee: work for hire law and its three-part scope of employment analysis. Gathering virtually every reported case, this section shows that work for hire law is more balanced than invention assignment law in weighing the interests of employers and employees.

With this divergence in mind, Section IV explains how both of these categories of intellectual property law can overlap in the same employee-created material. This is where the conflict between these doctrines arises – where applying one test instead of the other could lead to a different outcome in a dispute over whether the employer or the employee owns what the employee has created. This section also provides a tentative solution. Where the employee owns the copyright in a work under the work for hire doctrine, and therefore has a right to publish the work, the employer’s competing invention assignment claim should fail, because it is based on trade secrecy. As the Fourth Circuit held in a 1994 decision,[2] a decision which may provide the key to resolving the latent conflict between these areas of law, a competing trade secret claim cannot exist where a copyright-owner has the right to publish (and thus to disclose) the work at issue.

Finally, Section V examines the public policy implications of the clash between invention assignment law and work for hire law. In the context where an employee operating in good faith creates material outside the workplace that has only an attenuated connection to the employer’s business, the more balanced regime seen under the work for hire doctrine provides better protection for employees forming new ventures and thus better promotes innovation. Asserting the work for hire doctrine where copyrightable material is at stake may prevent overreaching by employers where their ownership claims are weakest.

Employees around the country are subject to contracts which grant their employers ownership in workplace creations – and sometimes also outside-the-workplace creations. Given the importance of these questions to mobile employees and those who wish to form new ventures, this essay looks to ask new questions about this corner of the law.

I. Setting the Stage: Pre-Departure Creative Activities

This essay focuses on a narrow period of time and the intellectual property laws that govern that moment. To set the stage for highlighting a conflict between two areas of intellectual property law, we focus on the point when an employee begins to think about changing jobs, perhaps starting his or her own new venture, or when an employee simply creates work on the side, for purposes not yet known. We consider the situation in which the employee does not misuse the employer’s trade secrets, the employee works on the proverbial nights and weekends outside the office, and the employee is not attempting to create material that is related to the employer’s business.[3]

Many people who would like to quit and start a new venture understandably do not want to resign until a concrete plan is more or less in place. Sometimes that means weeks or months of planning with other nascent co-founders or slowly seeking funding. If the employee creates no intellectual property before leaving, the only legal questions at stake are those of fiduciary duty, the duty of loyalty, and making preparations for the next job.[4]

Sometimes, however, pre-departure planning also means that one or more employees creates new ideas, and commits those ideas to writing or to software code. Assuming the ideas or expression are original, novel, or not generally known in the relevant industry, he or she has probably created some form of intellectual property. So the question arises: who owns that intellectual property – the current employer or the employee who created it?

Events such as these are hard to capture for analysis. They occur in secret, because the employee fears that discovery will result in immediate termination.[5] Further, where attorneys become involved in counseling departing employees, the attorney-client privilege means that practitioners are not free to discuss particular situations in a public setting. Even where the employer later discovers the pre-departure preparations and sends a cease-and-desist letter, many potential lawsuits are averted. Probably only a small fraction of such cases result in lawsuits, and even fewer court rulings are available in public databases.

To delve into the ownership disputes that may arise in these contexts, we will put aside disputes where departing employees misuse employer trade secrets, or create work product that overtly relates to (or even competes directly with) the business of the employer. Our focus will be situations where the employee intends to launch a different, non-overlapping business, and creates new intellectual property with that end in mind. That is, we will assume the departing employee’s good faith, and analyze who might own such work product when both the Copyright Act and the law of invention assignment contracts could encompass the work at issue. We begin with invention assignment law. [6]

II. Invention Assignment Contracts and Employee-Employer Conflicts

In order to understand how invention assignment law and work for hire law are sometimes in conflict, we will first examine the historical development of invention assignment law and then turn to the body of case law, which tends to significantly favor the employer over the employee.

Invention assignment agreements are commonplace in employment contracts. They reverse the historical common law presumption that an employee owns what he or she invents or develops, and they broaden a longstanding common law exception, which allows for employer ownership when the employee was hired to invent the type of intellectual property he or she then created.[7] Terms in invention assignment contracts address ownership of trade secrets and patentable inventions, but not copyrights.[8] Although these agreements allow for employee ownership in certain contexts, by and large they substantially favor employers.[9]

While it is hard to find a definitive source pinpointing when employers began using written invention assignment contracts to alter default common law ownership rules, the practice certainly dates back many decades. In one 1938 California case, for example, a court stated that “[a]n employee’s agreement in the contract of employment to assign patents to his employer is specifically enforceable as to patents, clearly within its terms, as strictly construed against the employer.”[10]

A. State Invention Assignment Statutes and Default Common Law Rules

In most states, invention assignment contracts are governed by common law, though there are not many reported cases in jurisdictions outside of California. However, eight states, beginning with Minnesota in 1977, have enacted similar statutes that seek to lay out boundaries for when employers can obtain ownership of employee creations, all of which favor the employer.[11]

California’s version is illustrative.[12] While awkwardly phrased, it provides six different ways in which the employer can own the employee’s creation: (1) use of the employer’s trade secrets; (2) use of the employer’s time; (3) use of the employer’s resources; (4) the creation is related to the employer’s current business; (5) the creation is related to the employer’s “demonstrably anticipated” research and development; and (6) the creation results from the employee’s work for the employer.[13] These possibilities are disjunctive, meaning that the employer need only demonstrate one of them to prevail.

State statutes regulating invention assignment agreements are not entirely identical. For example, California’s version requires that employers provide notice to the employee, and also allows employers to require that employees disclose all new creations – even if outside the scope of contractual ownership terms – for an ownership review “in confidence.”[14] Washington does not require that the employer treat employee disclosures in confidence, but gives the employee a right to disclose the creation to “the department of employment security.”[15] Both California and Washington put the burden of proof on the employee to show that the creation is outside the scope of employer ownership.[16]

A ninth state, Nevada, enacted a statute relating to invention assignment agreements later than the states discussed above, in 2001, with a different aim.[17] In what was apparently seen as a bid to improve the business environment for technology companies in the state, the statute renders an invention assignment contract unnecessary and broadens the scope of employer ownership compared to the common law. It states: “Except as otherwise provided by express written agreement, an employer is the sole owner of any patentable invention or trade secret developed by his or her employee during the course and scope of the employment that relates directly to work performed during the course and scope of the employment.”[18] As one commentator noted at the time, this sort of one-sided regime is unlikely to foster the birth of new technology ventures in Nevada.[19]

State law invention assignment agreements have been the subject of a fair amount of commentary, albeit mostly limited to the question of whether the default rules granting employers ownership of workplace inventions fairly compensate employee-inventors in a way that best promotes innovation equitably.[20] Some commentators have addressed other issues about the reach of invention assignment agreements which provide helpful background for the questions discussed in this article.[21]

Contracts between employers and employees dictating ownership of workplace creations arose because the default common law rule provided a narrower path to employer ownership. Specifically, case law dating to the nineteenth century provides that an employee owns his or her inventions, even when it is conceived during the term of employment,[22] subject to an important but narrow exception. Those who were “hired to invent” the same type of work product they then duly created owed an automatic assignment duty to the employer, vesting title in the company. These rules developed from early Supreme Court decisions.[23] In turn, the common law provided that if an employee who was not hired to invent nonetheless created an invention using the employer’s time or resources, the employee retained ownership, but the employer received a permanent, royalty-free license known as a “shop right” to practice the invention. Historical case law on the hired-to-invent exception and the shop right rule – that is, before the age of ubiquitous written invention assignment agreements and state statutes governing them – were often fact-intensive.[24] Even in more recent cases, these common law rules still apply in situations where, for whatever reason, there was no written contract.[25]

B. How Invention Assignment Contracts Favor the Employer

By and large, invention assignment contracts and related state statutes favor the employer over the employee, even though the text of some such statutes appears, at first glance, to be designed to protect employees. Generally speaking, an effective invention assignment to the employer occurs automatically, with no further steps required, and applies to inventions that the employee has not yet conceived.[26] In some jurisdictions, the contract can be effective even if the employer requires the employee to sign it after the employee has commenced employment.[27]

Expansive invention assignment statutes and agreements pose risks for employees who create outside-the-workplace developments, often with an eye towards forming a new company after resigning. Even if we assume good faith, with no use of employer trade secrets, time, or equipment, the employee faces many risks, and lines are not clearly drawn.

Part of the complication stems from changes in the workplace. When state invention assignment statutes were enacted in the 1970s and 1980s, perhaps it was easier to imagine a bright-line separation between the workplace and “nights and weekends” in an employee’s garage. Without the internet, employees typically did not work from home, much less on employee-owned phones and laptops. Moreover, it may then have been easier to determine with certainty what “related to” the employer’s business and what did not. Fewer companies were large multinational entities, with unrelated departments and operations of which an employee might be unaware. Though the scarce case law is hardly determinative on these questions, changes in modes of employment and the scope of companies’ operations may work to the detriment of employees who create new material on the side.

For similar reasons, employers may be able to argue that working hours – and thus what constitutes the employer’s time – are elastic, even 24/7, if employees typically answer email at home or on weekends. In one case that focused primarily on fiduciary duty but also involved an invention assignment contract, a court which issued a preliminary injunction was amenable to the notion that employer time meant more than ordinary working hours where employees often put in long days.[28] It remains to be seen how far this will go, however, as it seems that no court has yet analyzed whether a statute which appears to divide employer time and off-hours could be rendered meaningless if the employer is permitted to define its time as “24/7/365.”

A similar interpretive issue arises with the definition of the employer’s resources. While employees of technology companies in 1980 or 1990 used office workstations too unwieldy to take home, employees today commonly use company-issued laptops and mobile devices. Confusing the issues further, many companies permit dual use. For example, employees may use company-issued devices for personal use (or vice-versa), subject to codes of ethics or other workplace guidelines. If an employee creates an unrelated invention for a future venture on a weekend, but uses a dual use company-issued laptop to do so, can the employer claim ownership? If an employee creates a new idea for a future venture on a dual-use company-owned mobile phone, can the employer claim ownership? The answer again is unclear, as courts do not appear to have yet analyzed this question.[29]

The ambiguity of terms that appear in invention assignment statutes and contracts – such as terms that vest ownership in the employer when the employee’s creation relates to the employer’s business or research or when it results from the employee’s work – also leads to interpretive difficulties.[30] These terms notably alter the common law, under which an employee must be “hired to invent” in order for the employer to own the creation. Under these terms, even the receptionist would be subject to employer ownership if his or her outside-the-workplace creations could be said to relate in some fashion to the employer’s business.

More importantly, it remains unclear what phrases like “relates to” or “results from” mean in the everyday workplace. For example, what happens if an employee in good faith creates something offsite, intending to use it for a future business, but unbeknownst to the employee the company has been dabbling in something similar at a different office on the other side of the country? In other words, are the rules to be interpreted in a manner that takes the employee’s notice of the scope of the employer’s business into account? Does an employee’s good faith intent matter? Thus far, no reported case has addressed this question.

The closest may be a 2007 decision from the Northern District of California which ruled in favor of the employer where an employee argued that he should own his invention because his own department’s business did not encompass the work. The court, however, found that California’s invention assignment statute did not contain terms limiting it to the particular department in which the employee works.[31] That said, it appears that the employee was on notice that his or her invention related to the employer’s business,[32] and thus the question of whether the statute is to be interpreted to reflect the employee’s notice or state of mind is not at issue.

As explained below, the work for hire rules do explicitly consider the employee’s state of mind, setting up a potential conflict with the law of invention assignment contracts on these issues.

Further complicating the problems with the interpretation of invention assignment contracts, some companies define themselves in employment agreements to include their affiliates; different entities that are separately incorporated and separately managed. It is unclear whether an invention assignment statute or agreement could divest an employee of an outside-the-workplace invention in a situation where the employer contends that the invention relates only to the business of a differently-incorporated affiliate entity.

The expansive wording of invention assignment contracts and statutes is not the only difficulty facing employees who create offsite work product. Some courts have held that even where an employee notifies the employer about a development and the employer verbally disclaims any interest in it, the employer can later retract that apparent waiver. The reasoning is that since the invention assignment rules automatically transfer ownership at the moment of creation, if the employer owns the invention, there is nothing that can be verbally bargained over.[33] Because this sort of thinking is unlikely to occur to non-lawyers, this can pose an unfavorable trap for employees, to say the least.

Taking the concept of employer ownership to its extreme, some employers attempt to obtain ownership of developments created by former employees entirely after the employee has left the job. Known as “holdover” or “trailer” clauses, these terms typically assert ownership of inventions for a set period after the employee’s departure, or purport to establish a presumption of ownership. Some courts have rejected such clauses, at least where the former employee has not misappropriated the former employer’s trade secrets, or has narrowed their potential applicability.[34] But in 2012, the South Carolina Supreme Court found a broad holdover clause enforceable under the law of that state.[35] The court’s reasoning was similar to those of courts which enforce post-employment non-competition covenants. It also failed to analyze the case of former employees working in good faith to launch a new venture. The threat posed by holdover clauses to employees who resign and create new intellectual property for the first time after leaving is obvious.[36]

At the same time, courts have generally rejected employment contracts which purport to require employees to assign inventions created before starting the new job.[37] However, the employee’s incorporation of pre-existing work product into intellectual property created during the period of employment generally will not prevent assignment of that new work to the employer, assuming the conditions for such assignment exist.[38]

Finally, if an employee signs two invention assignment contracts, the first-in-time has priority. This means that if an employee has signed a valid invention assignment contract with a current employer, he or she cannot avoid that agreement by signing a new, second-in-time invention assignment contract in favor of a new start-up venture. Thus, creating an outside-the-workplace invention and assigning it under a new agreement to a new venture would not transfer ownership to the new venture if the prior first-in-time assignment agreement applies to vest the employer with ownership.[39]

C. The Case Law on Invention Assignment Contracts is One-Sided

Overall, these rules tend to favor the employer where an employee creates an outside-the-workplace invention. Where exactly an invention stops “relating to” or “resulting from” the company’s business or the employee’s work is not clear. That ambiguity empowers employers over employees. Adding further ambiguity about what constitutes use of the employer’s time or equipment, and the open question whether the employee’s personal notice of the employer’s activities matter, the statutes pose a minefield for offsite creations. Oral permissions may not suffice, undermining above-board efforts to be transparent. And in the most restrictive jurisdictions, former employers may even be able to obtain ownership of intellectual property conceived and developed entirely after the employee has quit and left.

Given the reach of state statutes and the manner in which these agreements override the common law “hired to invent” exception to broaden the scope of employer ownership, it is not surprising that the available case law strongly favors employers in disputes with employees, and in fact is nearly one-sided. Employees rarely prevail.[40]

Even a rare exception – a remand ruling in favor of a departing employee – rests on dubious conclusions. In the well-publicized Mattel v. MGA battle over ownership of doll designs where a departing employee had ideas and did at least some preliminary work before resigning, the Ninth Circuit found fault with the employer’s position because its invention assignment contract lacked the word “ideas” in its recitation of what the contract covered.[41] That ruling comes across as the product of incomplete analysis given that the contract seemingly included synonyms (such as “developments”) and because the court did not analyze the terms of California’s invention assignment statute. [42]

All told, and excluding cases about post-employment “holdover” clauses, there are many cases where the employer (whether a private company or a university) prevailed in an invention assignment contract dispute with an employee.[43] By contrast, only a few have found in favor of an employee.[44]

But as discussed below, a different regime governs when employees create copyrightable works – with different considerations, different outcomes, and potential direct conflicts with the law of invention assignment agreements.

III. The Work for Hire Doctrine and Employee-Employer Conflicts

A. Work for Hire Law and the Restatement of Agency

This article proposes that where an employee creates copyrightable material, the test for allocating ownership of that material differs from the invention assignment tests for allocating ownership of trade secrets and patentable inventions described above. We start with an analysis of modern work for hire law, developed following a major Supreme Court decision in 1989, Community for Creative Non-Violence v. Reid.[45]

The law of employee invention assignment contracts is pervasive, and encompasses most of the creative work product employees may generate on the job – whether trade secrets or inventions – that result in issued patents. However, it does not cover everything, and thus a separate, but seemingly similar body of law fills the gap. For employee creations that are expressive – and that meet the standards for copyright protection – the work for hire doctrine allocates ownership between employers and employees, just as the interplay of contract wording and (in some states) legislative enactments allocates ownership of potential trade secrets and patentable inventions.

The work for hire doctrine allocates ownership of copyrightable works in a manner that is more favorable to the employee, posing a direct conflict with the law of invention assignment contracts where an employee creates something that is potentially both copyrightable and a trade secret.

The work for hire doctrine is partly a matter of statute, partly a matter of contract law, and partly a matter of the common law of agency. Generally speaking, a copyrightable work created by an employee within the scope of his or her employment belongs, at creation, to the employer:

A “work made for hire” is (1) a work prepared by an employee within the scope of his or her employment; or (2) [specific categories of works not applicable to this essay].
In the case of a work made for hire, the employer or other person for whom the work was prepared is considered the author for purposes of this title, and, unless the parties have expressly agreed otherwise in a written instrument signed by them, owns all of the rights comprised in the copyright.

See 17 U.S.C. §§ 101, 201(b) (2012).[46] The work for hire doctrine does not extend to independent contractors, however, and thus a copyrightable work created by an independent contractor belongs to the contractor, absent a contract saying otherwise.[47]

In determining who is an employee and who is not, courts are to apply the federal common law of agency, rather than the employment law of particular states.[48] That question is solved by analyzing a large number of factors.[49] Disputes over whether a person is a contractor or an employee are common, and the case law on that question is extensive. However, where the individual at issue starts as a contractor and later becomes an employee, the question of whether a work for hire exists is further muddled.[50]

For our purposes, these disputes are peripheral. What matters instead is who owns the copyrightable material created by an employee when an employment relationship exists. As the Copyright Act states, when the author of a copyrightable work is clearly an employee of the party claiming ownership under the work for hire doctrine, the sole question is whether the author created the work within the scope of his or her employment:

A “work made for hire” is . . . a work prepared by an employee within the scope of his or her employment.[51]

Much like the law of invention assignment contracts, an employer’s agreement with an employee to assign works for hire is not the end of the story. A penumbra of case law defines what “scope of his or her employment” means, and in effect creates exceptions to the presumption of employer ownership.[52] Thus, while employment contracts ubiquitously recite that works for hire belong to the employer, the assignment is not monolithic.[53]

Every court to address the “scope of employment” issue since the Supreme Court addressed work for hire law in the 1989 Reid decision has applied a three-part test based on section 228 of the Restatement (Second) of Agency: “Conduct of a servant is within the scope of employment if, but only if: (a) it is of the kind he is employed to perform; (b) it occurs substantially within the authorized time and space limits; (c) it is actuated, at least in part, by a purpose to serve the master.”[54] This reference to the Restatement is typically perfunctory, as courts do not cite other sections from the same chapter.[55]

Supposedly, all three tests must be met for there to be a work for hire, meaning that the test is conjunctive, not disjunctive.[56] But at the same time, courts seem to place the least emphasis on the second factor and will find a work for hire if the other two factors are met and at least some of the creative work was done at the employer’s facility and during work hours – or in some cases none at all.[57] Further, like invention assignment law, courts may be open to an expansive interpretation of employer time where a company has a more 24/7 working culture.[58]

Perhaps because the work for hire scope of employment test presents a more balanced allocation of interests compared to the tests under a typical employee invention assignment contract, case law outcomes appear less one-sided. Putting aside inconclusive rulings, I have found around the same number of work for hire cases where the employer prevailed as to ownership of the work (at least as to injunctive relief)[59] as those where the employee prevailed.[60]

Scholarly writing about the work for hire doctrine has mostly centered on a question that precedes the topic of this article – whether the author is an independent contractor or an employee in the first place.[61] That focus is no surprise, as there is a prominent Supreme Court ruling (Reid) on that issue. Moreover, the rise of the so-called “gig economy,” where companies treat workers as independent contractors rather than employees, may also explain why many focus on the rights of contractors versus those who hire them. By contrast, commentary on the scope of employment test has been meager, with some exceptions noted below. The same is true in the leading treatise on copyright law.[62]

That said, several articles provide useful background for the issues analyzed here. A student note comes perhaps closest to the issues presented here, by contrasting copyright and patent rights in employee-created software.[63] Another commentator, in a detailed overview of the allocation of ownership interests in works created by employees, advocates for “an interpretive rule operating in favor of the employee” given that “[e]mployers are likely to attempt broad and vague definitions of the job description” in order to satisfy the first factor of the scope of employment test.[64] Finally, a definitive history of the work for hire concept and the transition to employer ownership of employee work product sheds light on the origins of the conundrum this essay addresses.[65]

B. The Scope of Employment Case Law is More Balanced

The manner in which courts have analyzed the three-part scope of employment test in work for hire disputes is noticeably different from the analysis seen in invention assignment disputes. As described in the cases summarized here, courts tend to focus most on the first and third factors of that test. And in contrast to invention assignment agreement law, employees prevail more often because the three factors are more employee-friendly.

Case Law on Factor One (Hired to Perform): So long as the employee’s job description/job duties are of the same general category as the copyrightable work, employers tend to win on the first factor – even if the work is only incidentally related to the employee’s job duties. The employer’s provision of guidance and supervision to the employee weighs in the employer’s favor as well.[66] At the same time (and even if the work might be incidentally related to the scope of employment) the employer’s disclaimer of the work is highly relevant, in contrast to the law of invention assignment contracts.[67] Conversely, if the employee is hired to do tasks in categories unrelated to the copyrightable work, the employee tends to prevail on the first scope of employment factor.[68]

Case Law on Factor Two (Employer’s Time and Space): Much of the case law on the second scope of employment factor focuses on whether or not the work was done at the employer’s facility, though the test as worded is broader than that inquiry. The focus is where the work was created, not where it was later used.[69] If job duties necessarily include working off-site, the fact that an employee creates the work at home may still favor the employer.[70] However, in many cases, the employee’s work off-site means that the employee wins this factor.[71]

That said, prevailing on the second factor may not favor the employee overall if the work was done during the time the employee was employed. Courts may not find the second factor determinative if the employer has prevailed on the first and third factors.[72]

Case Law on Factor Three (Employee’s Purpose): Sometimes there are cases where the employee did not seek to benefit the employer, beyond any doubt.[73] And sometimes the employee’s copyrightable work is developed in tandem with work done for the employer, so that it cannot reasonably be anything other than a work for hire.[74]

However, most cases are closer. Many involve a situation where an employee created something at home, used it at work but also for other purposes, and was motivated by both personal motives and a desire to help the employer. As noted above, the second factor can be discounted if the other factors are in the employer’s favor. That makes sense, as otherwise an employee could create something at home and automatically get ownership of it, no matter how closely it relates to his or her job role.

So where do courts draw the line when the employee may have had mixed motives? In general, the motivation to assist the employer need only be partial; it need not be the sole motivation. The employee must be at least “appreciably” motivated by a desire to serve the employer’s goals.[75] But what does this mean? Greater than a 50 percent motivation to serve those goals? Something else? The cases do not nail it down with precision. Both the employer[76] and the employee[77] can prevail on this factor.

Some courts, such as the Beasley and Roeslin cases cited below, suggest that looking at the employee’s subjective beliefs is one approach to analyzing the third scope of employment factor. One court made this approach the dispositive test. In Martin v. City of Indianapolis, the court, relying on the Restatement (Second) of Agency section 235, comment (a), stated that “[t]he state of mind of the employee is the material determination; the court may consider the employee’s actions or other manifestations only as evidence of the employee’s state of mind.”[78] Under that sort of analysis, the employee won easily, because he claimed that his sculpture was a personal project, that he did not get paid for it, spent a lot of time on it, worked on it in his free time, signed an agreement calling himself the “owner” of it, and was, apart from his work for the employer, “a serious, independent artist who uses stainless steel as a medium.” Focusing on the employer’s potential economic motivations, the court noted that the work was not created in order for the employer to get a commercial benefit, as it never sold the work to a customer.[79]

Another way to consider the question is to look to whether the employee was doing the work for a third party at the time – as doing the work for someone else is surely relevant to whether the work was done “appreciably” for the employer. In one case, the work in question (a bag design) was of the type the employee was hired to do, and the first factor favored the employer. However, the work was not a work for hire, because the employee created the design for a third party, assigned it to that third party, and was paid directly by the third party. Thus, the design was not “actuated by a purpose to serve [the employer’s] interest.”[80]

Overall, an employer can win on the first factor but still lose the third, and the focus – however a court words it – appears to be whether some significant percentage of the employee’s subjective motive was to help the employer, focusing on the time the work was created.[81] Because the law of invention assignment contracts does not expressly consider the employee’s intent when creating intellectual property, this consideration of the employee’s state of mind is quite different from the analyses seen under that body of law.

IV. Is Work for Hire Law Inconsistent with Invention Assignment Law?

This review of the work for hire doctrine and the law of invention assignment agreements shows why there is seemingly a direct conflict between them. Where an employee creates something that is both copyrightable and that could be an employer-assigned trade secret, the outcome of an ownership dispute could differ.

Unlike trade secrets and patent rights, trade secrets and copyrightable works are categories of intellectual property that can exist concurrently in the same work. Indeed, because the Copyright Act does not preempt state trade secret law, courts deem that confidentiality requirements sufficiently distinguish a state law trade secret cause of action from a copyright infringement cause of action. A plaintiff with ownership rights under both can theoretically pursue both causes of action against a defendant regarding the same work.[82]

Thus, it follows that when an employee creates new work, that work might constitute the employer’s trade secret – if it meets the conditions for automatic invention assignment under the employment agreement – and it also might constitute the employer’s work for hire – again, if the conditions for the scope of employment test are satisfied. In many cases, especially for routine situations where the employee creates work for the employer in the ordinary course of business, both tests will easily be satisfied, and nobody would imagine otherwise.[83]

However, on the margins, where employees create work outside the workplace and for outside purposes, the two tests can dictate different results as to the same work by the same employee. To begin with, the invention assignment analysis is disjunctive – the employer need only satisfy one of the conditions for ownership, such as demonstrating that its equipment was used. By contrast, the work for hire analysis is conjunctive – the employer must show all three factors, or at least such a strong showing on the first and third that the second is discounted.

Second, the invention assignment analysis is agnostic to an employee’s job role. If the receptionist writes software that relates to the employer’s business, and if the receptionist signed an invention assignment contract, the employer owns the work even if the receptionist programmed it during off-hours and away from the office. The invention assignment contract thus reaches beyond the common law “hired to invent” test for employee ownership, as discussed above. Again, by contrast, the work for hire scope of employment test looks to the employee’s job role as the first factor of the analysis, thereby providing a narrower path to employer ownership.

Perhaps most importantly, the invention assignment analysis is agnostic to the employee’s motive or intent. Even if an employee creates an off-site work in anticipation of a new job or creating a new company, that motive is not relevant if the work is found to relate to the employer’s business or relate to its demonstrably anticipated research and development.[84] But as discussed above, the third factor of the work for hire scope of employment test significantly considers the employee’s motive.

With these differences in mind, there are many decisions where the outcome might have differed depending on which of the two doctrines the court focused on (or perhaps more accurately, the doctrine the employer focused on when bringing the lawsuit).

One example where a conflict might have arisen had the court spent more time on the employer’s invention assignment claim was a 2004 ruling from the Southern District of Iowa, PFS Distribution Co. v. Raduechel. There, the employee was a truck driver for a poultry plant. A self-taught programmer, he created software at home for an order entry system. The employee licensed his software to two unrelated businesses and later created a version for use by his employer, brought it on site and installed it, and tailored it for the employer’s business.[85]

After the company was acquired and cut salaries, the employee and a co-worker planned a mass departure for a new venture, engaging in significant misconduct that led to a preliminary injunction.[86] However, the employer’s claim to ownership of the truck driver’s order entry software faltered because the court agreed that the employee owned the work under the work for hire doctrine. He was not hired to program, he created the work at home and even licensed it to others before bringing it to work, and his purpose in creating it was not to serve the employer.[87]

Raduechel is possibly the only case where a court considered a work for hire dispute alongside a claim for breach of an invention assignment contract. Had the court spent more time analyzing the latter, it might have recognized a potential conflict. After all, the employee’s software surely related to the employer’s business, and at least a version of it was modified for use at the employer’s facility and presumably on its equipment. Under the employer-friendly regime of invention assignment law, perhaps the court might have found that the employer owned at least a portion of the work. However, the court appears to have brushed off the employer’s invention assignment cause of action, having already decided the work for hire issue and quickly concluding that the result was the same under the employer’s contract claim.[88]

Another case illustrating the potential conflict is a 2010 ruling from the Central District of California, TAP Worldwide, LLC v. Becker. There, the employer sold “off-road parts and accessories.” The employee prevailed in a work for hire dispute over ownership of a software program he wrote to expedite the processing of export shipments.[89] He had created the work on his own time, and based on his own “market research,” without using the employer’s equipment.[90] The employer sought ownership and an order requiring the former employee to “restore access” to the software, but lost on all three work for hire scope of employment factors.[91] Would the result have been the same if the employer had sued for breach of an invention assignment agreement? In California, it would need only show that the work related to its business, and not that the employee was hired to create that type of work, or that the employee was motivated by a purpose to serve the master.[92]

Perhaps the case that comes closest to demonstrating the conflict between the work for hire doctrine and invention assignment law is a 1994 Fourth Circuit case where an employee’s ownership of the copyright in a work precluded the employer’s dueling claim to trade secret rights in the same information. In Avtec Systems, Inc. v. Peiffer, the employer provided simulations of orbital patterns to the federal government. The employee, whose job duties included computer simulations for satellite orbits, wrote a software program for that purpose, but did so on his own time. The company used the program as a marketing tool, and other employees suggested modifications. However, the employee later licensed the program to a third party.[93]

After a confusing bench trial ruling where the lower court found that the employee owned the copyright but the employer held a “shop right” and/or a “trade secret” in the same work, the Fourth Circuit remanded for clearer findings on the work for hire question. Crucially, the court rejected the notion that an employee’s copyright ownership could be undermined by an employer’s trade secret claim to the same information. Noting that copyright ownership confers a right of publication and distribution, while trade secret rights depend on non-publication and secrecy, the court held that “Avtec offers no authority, and we have found none, for the proposition that the alleged ‘owner’ of a trade secret . . . could maintain the secrecy of material that is subject under federal law to publication at the will of another.”[94]

Ultimately, after remand, the Fourth Circuit affirmed a ruling that the employee owned the copyright under the work for hire factors, and confirmed that an employee-owner of a copyrighted work cannot be subject to an employer’s trade secret misappropriation claim over that work.[95]

Although the employer brought trade secret and fiduciary duty causes of action against the employee rather than a breach of contract claim relating to invention assignment, that is a distinction without a difference for our purposes, as the intellectual property rights arising from an invention assignment contract would be trade secret rights. Thus, Avtec teaches that, should an employee prevail on copyright ownership under the work for hire analysis, an employer likely could not succeed in a breach of invention assignment claim over trade secret rights in the same work, as the conflicting trade secret claim would be void ab initio.[96]

Given this under-analyzed conflict, an employee’s potential copyright ownership of an outside-the-workplace creation may well be the dispositive question in clashes where an employee prepares work product for an outside opportunity while still employed. Litigants’ choices about what law to focus on during ownership disputes over copyrightable works may be more important than has been recognized thus far.

Conclusion – The Use-Value of the Scope of Employment Test

This essay is not about ownership of ordinary-course creative work performed in the workplace on behalf of an employer. It is instead about marginal cases where an employee, operating in good faith, avoids the employer’s trade secrets, time, and equipment to start preparing for a new job, forming a new company, or simply creating for personal interest. That is the context where employer ownership is most debatable, and that is the context where the clash between the work for hire doctrine and invention assignment agreement law is sharpest.

As we have seen, it is not difficult to imagine a situation where an employee creates something outside the workplace that is both copyrightable and something an employer might claim as its trade secret through contractual assignment. Whereas the scope of employment analysis might provide that the employee was not hired to create the sort of work in question, did not create it with the employer’s time or resources, and did not create it to serve the employer, invention assignment agreement law might find that the work nonetheless can be deemed “related” to the employer’s business in some manner.

Heightening the risk for employees, these rules are not intuitive. The work for hire doctrine is barely a ripple in the ocean of intellectual property law. Reported invention assignment disputes over outside-the-workplace creations that are not clearly related to the employer’s business are rare. Employees may not recognize that an employer may stretch invention assignment law to find ways to argue that the work relates to the employer’s business in some unexpected fashion. And as discussed above, if the employee were unaware of that purported connection between his or her at-home work and some project contemplated by a far-flung, international employer, it is not clear that the employee’s lack of notice would play a role in the analysis. In jurisdictions like South Carolina, invention assignment terms can reach beyond the grave, as it were, to allow past employers to claim ownership in work product created long after an employee’s departure.

This potential for overreach endangers employee mobility and subjects new enterprises to aggressive legal claims by former employers. Many employees may begin creating work product for a new venture before leaving their current job. Some will break the rules in clear-cut ways, but others try to do the right thing, yet still face legal traps for the unwary. Not everyone can afford to quit and then start development work. And not everyone can afford legal advice before quitting.

How should such conflicts be resolved? And how can these issues be resolved in a manner fair to an employee operating in good faith? As to copyrightable work, one answer is that the work for hire rules should control over the trade secret-centered rules of invention assignment agreement law. That is not merely because copyright law is federal law while invention assignment law is largely a matter of state contract and state statutory law.[97] After all, trade secret rights can now exist under federal law as well, and some aspects of invention assignment contracts are also matters of federal law.

Rather, the work for hire rules control because copyright ownership precludes a conflicting claim to trade secret ownership, whether via an invention assignment contract or otherwise. As the Avtec case explained, if an employee owns the copyright – and thus the right to disclose and publish the work – an employer’s conflicting trade secret claim is not possible because there can be no trade secret rights in something the adverse party has the freedom to disclose at will. To put it in the language of invention assignment law, an employee cannot agree to a future conveyance of a trade secret right that would be stillborn at the outset given overriding copyright ownership under the work for hire doctrine’s scope of employment test.

This outcome is the better result for outside-the-workplace creations where the employer’s ownership claim would require an attenuated argument that the employee’s at-home work “relates” to the employer’s business in some strained or unexpected manner – or where the employee had no notice that the employer was thinking of something similar in another office, or with a different engineering team. Where work product is copyrightable, the scope of employment test under the work for hire doctrine could become a tool to prevent overreaching, and thereby to protect departing employees.

To be sure, this is hardly the same as full reform of a confusing regime, as this prospect only applies in cases where the employee creates something copyrightable. People changing jobs face a confusing mix of potential claims, and this is only part of the difficulties they face.[98] Future essays in this series will examine others.


* Partner, Wilson Sonsini Goodrich & Rosati, San Francisco, and Adjunct Faculty, University of California, Hastings College of the Law.

** This essay is the first in a three-part series addressing under-analyzed areas of intellectual property and employee mobility law which impact creative employees when changing jobs. Departing employees can face a tangled body of contract, tort, and statutory claims brought by former employers. Academics and practitioners have provided little commentary about some of these overlooked areas of law. This relative inattention is surprising given the important policy concerns so often at stake in mobility disputes.

[1] One powerful, if rare, example of the sort of thoroughgoing critique that is all too often lacking is found in Orly Lobel, The New Cognitive Property: Human Capital Law and the Reach of Intellectual Property, 93 Tex. L. Rev. 789, 792-93, 798-99, 801-03, 813-15 (2015) (tying the overbreadth of employee invention assignment agreements – noting, among other things, post-employment holdover clauses – to broader concerns over employer overreach and resulting harms to mobility and growth of new enterprises).

[2] Avtec Sys. v. Peiffer, 21 F.3d 568 (4th Cir. 1994).

[3] In general, and while definitions vary slightly from state to state, a trade secret is non-public business information valuable to competitors that the owner has protected with reasonable security measures. E.g., Defend Trade Secrets Act of 2016 § 2(b)(1), Pub. L. No. 114-153, 18 U.S.C. § 1839(3), 130 Stat. 376, 380.

[4] Disputes over these so-called “preparations to compete” will be the subject of the next essay in this series.

[5] Even if preparations to compete are legal, an employer can terminate an employee found to be engaging in such activities. See Fowler v. Varian Assocs., Inc., 241 Cal. Rptr. 539, 544 (Ct. App. 1987) (granting summary judgment for employer where employee brought wrongful termination claim after being fired for making preparations to leave and compete).

[6] Readers may wonder why this essay refers to employers’ trade secret rights through invention assignment agreements, rather than patent rights. After all, and as the very phrase “invention assignment” suggests, patents are what scholars and practitioners likely think of first, if not exclusively, when the subject of invention assignment arises. But this misses an important timing point: when an employee conceives a patentable invention, it first exists as intellectual property in the form of a trade secret, not a patent. If the employer chooses to pursue a patent application, that election comes after the point of initial, automatic assignment. That is why this essay points to a clash between competing copyright and trade secret rights in the same work product.

[7] While ownership of an invention would otherwise vest in the inventor, the standard rule and practice is that an employee can contract ownership away through an invention assignment agreement. See Beech Aircraft Corp. v. EDO Corp., 990 F.2d 1237, 1248 (Fed. Cir. 1993); Access Cardiosystems, Inc. v. Fincke (In re Access Cardiosystems, Inc.), 340 B.R. 127, 147 (Bankr. D. Mass. 2006) (explaining rules). Thus, “an employer owns an employee’s invention if the employee is a party to an express contract to that effect.” Banks v. Unisys Corp., 228 F.3d 1357, 1359 (Fed. Cir. 2000); Grove v. Grove Valve & Regulator Co., 84 Cal. Rptr. 300, 303-04 (Ct. App. 1970) (early case finding that employer owned improvements to various inventions in enforcing terms of employee invention assignment agreement signed in 1959).

[8] Although such contracts assign ownership of rights including those governed by federal law – i.e., patents, but also trade secrets given the 2016 Defend Trade Secrets Act – their interpretation is largely, but not entirely a matter of state law. Until 2008, the interpretation was solely one of state law. See Affymetrix, Inc. v. Illumina, Inc., 446 F. Supp. 2d 292, 296 (D. Del. 2006) (applying California law; “The proper construction of assignment agreements is a matter of state contract law.”) (citing Minco, Inc. v. Combustion Eng’g, Inc., 95 F.3d 1109, 1117 (Fed. Cir. 1996)). This default rule was modified in 2008, when the Federal Circuit ruled that, at least as to the question of whether an assignment contract creates an automatic future conveyance of inventions not yet in existence – which is accomplished by the inclusion of the magic words “hereby assign” – is one of federal law. See DDB Techs., L.L.C. v. MLB Advanced Media, L.P., 517 F.3d 1284, 1289-90 (Fed. Cir. 2008). DDB’s reasoning was based on federal courts’ powers as to “the question of standing in patent cases.” Id. See also Intellectual Ventures I LLC v. Erie. Indem. Co., 850 F.3d 1315, 1320 n.1 (Fed. Cir. 2017) (stating rule). The decision was notably silent on what law would govern the future conveyance of state law trade secret rights in an employee creation.

[9] See discussion and cases infra Section B.

[10] Hercules Glue Co., Ltd. v. Littooy, 76 P.2d 700, 701 (Cal. Dist. Ct. App. 1938) (enforcing employee invention assignment contract in favor of employer: “By the terms of this contract, appellant agreed that he would devote his entire time, knowledge, skill, best efforts, and services to the work of respondent as it might direct; that all patents, processes, or formulas, pertaining to spreader, which he invented, developed or perfected during his original and present employment, should be the property of respondent, and that he would, on demand, execute any assignments, transfers or other instruments necessary to perfect respondent’s title thereto. He further agreed that he would not, during its life or subsequently, divulge to any person, or use for his own benefit, any secret processes or formulas invented, developed, or perfected by him or respondent between October 1, 1924, and the termination of the agreement.”). Other early cases involving written contacts include: Kober v. United States, 170 F.2d 590, 592, 594 (4th Cir. 1948) (finding that an employment agreement with a U.S. government employer providing for “the complete assignment” of inventions made by the employee was a “reasonable agreement entered into for a lawful and proper purpose” and holding that inventions made by an engineer subject to this agreement therefore belonged to the government-employer); Crown Cork & Seal Co., v. Fankhanel, 49 F. Supp. 611, 613, 615 (D. Md. 1943) (construing and finding enforceable an invention assignment agreement that required the employee to assign rights over to the employer for “[a]ll inventions and discoveries which I make while in the employ of said Company, along the lines of its general work, constituting improvements both in its then existing products and methods of manufacture, or otherwise, shall become its exclusive property,” and holding that the employment agreement was valid and enforceable).

[11] See Cal. Lab. Code §§ 2870-2872 (West 1979); Del. Code. Ann. tit. 19, § 805 (1984); 765 Ill. Comp. Stat. 1060 / 2 (1983); Kan. Stat. Ann. 44-130 (West 1986); Minn. Stat. Ann. § 181.78 (West 1977); N.C. Gen. Stat. §§ 66-57.1-.2 (1981); Utah Code. Ann. § 34-39-3 (West 1989); Wash. Rev. Code Ann. §§ 49.44.140, .150 (West 1979). Based on this author’s experience, companies often mimic the structure of these state statutes in employment agreements, even for employees residing in other states. As a result, these eight statutes have great influence beyond their home jurisdictions.

[12] See Cal. Lab. Code §§ 2870-72 (West 1979).

[13] Id. § 2870. The “results from” clause is unclear, but its interpretation has not yet been the subject of a published decision.

[14] Id. § 2871.

[15] Wash. Rev. Code Ann. § 49.44.150 (West 1979).

[16] See id.; Lab. § 2872.

[17] See Nev. Rev. Stat. Ann. § 600.500 (West 2001).

[18] Id.

[19] See Mary LaFrance, Nevada’s Employee Inventions Statute: Novel, Nonobvious, and Patently Wrong, 3 Nev. L.J. 88 (2002) (well-done dissection of the then-new state statute, noting its departure from the common law, contrast with other state statutes, and contrast with the work for hire rules; predicting the statute would not, as apparently intended, encourage the growth of technology companies in Nevada).

[20] There is a cottage industry of notes and law review articles positing that if employees were given greater compensation for, or some ownership stake in, inventions created for the employer in the workplace, innovation would increase. See Shlomit Yanisky Ravid, Rethinking Innovation and Productivity Within the Workplace Amidst Economic Uncertainty, 24 Fordham Intell. Prop. Media & Ent. L.J. 143 (2013) (arguing that employees would have a greater incentive to invent if they retained a greater interest in – and received financial compensation for – workplace inventions); Ann Bartow, Inventors of the World, Unite! A Call for Collective Action by Employee-Inventors, 37 Santa Clara L. Rev., 673, 677 (1997) (advocating that employees refuse to sign invention assignment agreements and instead try to own workplace inventions themselves); Evelyn D. Pisegna-Cook, Ownership Rights of Employee Inventions: The Role of Preinvention Assignment Agreements and State Statutes, 2 U. Balt. Intell. Prop. L.J. 163, 185 (1994) (general survey of common law and state statutes; concluding by advocating additional compensation for employee-inventors and greater adoption of statutes to regulate invention assignment contracts); Steven Cherensky, Note, A Penny for Their Thoughts: Employee-Inventors, Preinvention Assignment Agreements, Property, and Personhood, 81 Cal. L. Rev. 595 (1993) (in an argument premised on a Hegelian-inflected notion of personhood and property rights, advocating a greater property allocation to employees for workplace inventions); Henrik D. Parker, Note, Reform for Rights of Employed Inventors, 57 S. Cal. L. Rev. 603, 624-25 (1984) (arguing that compensation schemes for workplace inventions would increase innovation); William P. Hovell, Note, Patent Ownership: An Employer’s Rights to His Employee’s Invention, 58 Notre Dame L. Rev. 863, 887-88 (1983) (advocating that employee-patent inventors receive a “reverse shop right” “to use any patent assigned to his employer”). All are speculative, and some rest on seemingly needless concepts of Hegelian personhood and the like. As much as I tend to favor departing employees, this line of articles is frustrating. Many seem to imagine that workplace inventions are created by a sole inventor rather than by teams, none contemplate the quagmires that could result in bickering over who owns what within the office, and the philosophical reference points seem arbitrary.

I consider all of these fully answered by Robert P. Merges, The Law and Economics of Employee Inventions, 13 Harv. J.L. & Tech. 1, 14-18 (1999) (arguing that the current employer-ownership regime is best calibrated to avoid problems such as disruptive employee “holdups”: “The employees could hold up the firm for the full value of its investment.”). Merges notes the seemingly obvious problems that would arise in scenarios where employees share in ownership of inventions made for the employer’s benefit. This essay is by contrast focused on outside-the-workplace creations by employees planning their next venture which have no connection, or only some tenuous connection, to their present employer’s business.

[21] Articles which address topics other than whether assigning workplace inventions to the employer is equitable include: Parker A. Howell, Whose Invention is it Anyway? Employee Invention-Assignment Agreements and Their Limits, 8 Wash. J.L. Tech. & Arts 79 (2012) (practitioner-oriented overview of state statutes and a sampling of case law); Joshua L. Simmons, Inventions Made for Hire, 2 N.Y.U. J. Intell. Prop. & Ent. L. 1, 49 (2012) (arguing that the law of employee-patent assignments contains too many procedural stumbling blocks and advocating that patent law be amended to match the work for hire terms of the Copyright Act); Donald J. Ying, A Comparative Study of the Treatment of Employee Inventions, Pre-Invention Assignment Agreements, and Software Rights, 10 U. Pa. J. Bus. & Emp. L. 763 (2008) (survey of invention assignment laws in the United States and other countries); Michael R. Mattioli, The Impact of Open Source on Pre-Invention Assignment Contracts, 9 U. Pa. J. Lab. & Emp. L. 207, 228-34 (2006) (article on what might happen if an employee creates a patentable invention while contributing to an open source project).

[22] See generally Solomons v. United States, 137 U.S. 342, 346 (1890) (“[W]hatever invention [an inventor] may thus conceive and perfect is his individual property.”); Gayler v. Wilder, 51 U.S. 477, 478–93 (1851) (“But the discoverer of a new and useful improvement is vested by law with an inchoate right to its exclusive use, which he may perfect and make absolute by proceeding in the manner which the law requires.” The case involved an assignment by the inventor to a businessperson.).

[23] “[U]nless there is an agreement to the contrary, an employer does not have rights in an invention ‘which is the original conception of the employee alone.’” Bd. of Trs. of the Leland Stanford Junior Univ. v. Roche Molecular Sys., 563 U.S. 776, 786 (2011) (quoting United States v. Dubilier Condenser Corp., 289 U.S. 178, 189 (1933)). “In most circumstances, an inventor must expressly grant his rights in an invention to his employer if the employer is to obtain those rights” because “the respective rights and obligations of employer and employee, touching an invention conceived by the latter, spring from the contract of employment.” Id. at 786. (quoting Dubilier, 289 U.S. at 187). Thus, the Court has “rejected the idea that mere employment is sufficient to vest title to an employee’s invention in the employer.” Id. at 789. However, there is an exception for employment situations where the employee has “only produced that which he was employed to invent.” Dubilier, 289 U.S. at 187. This “hired to invent” exception exists because the employee’s invention “is the precise subject of the contract of employment. A term of the agreement necessarily is that what he paid to produce belongs to his paymaster.” Id. at 187. “On the other hand, if the employment be general, albeit it cover a field of labor and effort in the performance of which the employee conceived the invention for which he obtained a patent, the contract is not so broadly construed as to require an assignment of the patent.” Id. The Supreme Court first applied the hired to invent exception in a 1924 case, Standard Parts Co. v. Peck. In that case, the defendant was employed under a contract to “devote his time to the development of a process and machinery for the production of the front spring now used on the product of the Ford Motor Company.” Standard Parts Co. v. Peck, 264 U.S. 52, 59 (1924). After inventing the specific process and machinery contemplated by his employment contract, the defendant secured a patent on the aforementioned technology which he refused to convey to the plaintiff. Finding that the invention of the specific patent at issue was the “object and effect” of the defendant’s employment contract, and that in inventing the technology at issue, the defendant was “doing nothing more than he was engaged to do and paid for doing,” the Court held that the patent belonged to the plaintiff. Id. at 59-60.

For other early cases, see Houghton v. United States, 23 F.2d 386, 389 (4th Cir. 1928) (Surgeon General appointed government employees to a board whose purpose was to develop a safer fumigant, and thus the defendant “did merely that which he was being paid his salary to do,” so there could be “no doubt” that his invention belonged to the government.); Magnetic Mfg. Co. v. Dings Magnetic Separator Co., 16 F.2d 739, 741 (7th Cir. 1926) (on contested facts, finding that employee was hired to invent from his own conduct, where he had promptly assigned two prior inventions, and thus owed an assignment duty even as to an invention that related to only a peripheral part of his employment duties). An employer might overcome hired-to-invent issues where a manager came up with the invention, and the employee claiming ownership was merely the one who reduced it to practice. In one early California case, the court distinguished between an employer who had the “general idea of a machine” and an employee who implemented the idea – even though the employee was the signatory on the patent application – in granting ownership to the employer in an improvement for movie studio lighting. Famous Players-Lasky Corp. v. Ewing, 194 P. 65, 66 (Cal. Dist. Ct. App. 1920).

[24] See, e.g., Fish v. Air-O-Fan Prods. Corp., 285 F.2d 208, 210-11 (9th Cir. 1960) (finding implied in fact assignment contract where evidence showed employee was hired to invent); Consol. Vultee Aircraft Corp. v. Maurice A. Garbell, Inc., 204 F.2d 946, 949 (9th Cir. 1953) (although employee had conceived invention before joining employer, he had used employer’s resources to perfect it, and thus employer had shop right to manufacture, use, and sell “airplanes embodying” the invention); Aero Bolt & Screw Co. v. Iaia, 5 Cal. Rptr. 53, 58 (Cal. Ct. App. 1960) (employee retained ownership of invention with no shop right for employer where, among other things, trial court found that employee “on his own time, and at his own expense, designed and developed the [invention]; that he paid all of the development costs out of his own pocket”); Zahler v. Columbia Pictures Corp., 4 Cal. Rptr. 612, 617 (Ct. App. 1960) (where employee was hired to create background music, court applied a version of the hired-to-invent doctrine to grant employer performance rights; “Where an employee creates something as part of his duties under his employment, the thing created is the property of his employer.”); Banner Metals, Inc. v. Lockwood, 3 Cal. Rptr. 421, 428-33 (Ct. App. 1960) (where salesperson had only an oral employment contract, court found that employer did not own the employee’s invention through the hired-to-invent-doctrine and also did not have a shop right in the invention, because the employee had created it on his own time and, on contested facts, had not used the employer’s materials); Quaker State Oil Ref. Co. v. Talbot, 174 A. 99, 101-04 (Pa. 1934) (employee was hired to invent under oral contract, but employer did not obtain shop right to employer’s pre-existing work that was not created with use of employer’s resources).

[25] See Peregrine Semiconductor Corp. v. RF Micro Devices, Inc., No. 3:12-CV-0911-H (WMC), 2014 U.S. Dist. LEXIS 2668, at *13 (S.D. Cal. Jan. 8, 2014) (where former employer sued former employee who transferred patent rights to a third party, and where employer could not prove it had a signed employment agreement, hired to invent doctrine did not favor employer on motion for preliminary injunction where it did not appear employee was hired for specific inventions at issue); Gen. Elec. Co. v. Wilkins, No. CV F 10-0674 LJO JLT, 2012 U.S. Dist. LEXIS 124642, at *59-60 (E.D. Cal. Aug. 31, 2012) (finding trial issues of fact on a hired-to-invent dispute where employee had not signed the company’s invention assignment agreement, but was a patent co-inventor); McClain v. State, 269 S.W.3d 191, 198-99 (Tex. App. 2008) (prosecution failed to show that defendant was hired to invent, such that improvements he made to “set up sheets” belonged to him, with a shop right to his employer); Pedersen v. Akona, 429 F. Supp. 2d 1130, 1143 (D. Minn. 2006) (in absence of invention assignment contract, finding that employer owned patent because employee was hired to invent); Scott Sys., Inc. v. Scott, 996 P.2d 775, 779 (Col. App. 2000) (reversing summary judgment where there were triable issues of fact as to whether individual was hired to invent or not with respect to certain inventions); Teets v. Chromalloy Gas Turbine Corp., 83 F.3d 403, 408 (Fed. Cir. 1996) (using the terminology of an implied-in-fact contract, but engaging in same analysis as hired to invent cases, reversing trial court and finding implied employee assignment obligation for invention where employee worked at company’s direction and used its resources); Liggett Grp., Inc. v. Sunas, 437 S.E.2d 674, 678-80 (N.C. Ct. App. 1993) (reversing grant of summary judgment in favor of employer and finding triable issues of fact as to whether employee with no invention assignment contract was hired to invent in connection with tobacco-related patent); California E. Lab., Inc. v. Gould, 896 F.2d 400, 402-03 (9th Cir. 1990) (defining shop right concept and finding that right transferred to successor entity); Aetna-Standard Eng’g Co. v. Rowland, 493 A.2d 1375, 1381-82 (Pa. Super. Ct. 1985) (employee with no written contract was not hired to invent and thus owned no assignment duty to employer of invention, but employer had shop right license because its resources were used in development); Mechmetals Corp. v. Telex Comput. Prods., Inc., 709 F.2d 1287, 1292-94 (9th Cir. 1983) (finding no shop right where company CEO had worked with two others to create invention, and company paid for materials and machine time CEO used when working on invention; existence of that contract and CEO’s status militated against finding shop right); Vigitron, Inc. v. Ferguson, 419 A.2d 1115, 1117-18 (N.H. 1980) (employer owned invention because employee was hired to invent).

[26] An invention assignment agreement with the phrase “hereby assign” (or similar language) operates to automatically vest a covered invention with the employer at the moment it is created by the employee. No further action – that is, no formal act of assignment by the employee – is required for the automatic invention assignment to be effective. See Roche, 563 U.S. at 785-93 (noting historical rule that employee-inventors maintain ownership of their inventions absent an agreement with the employer; university did not own invention where agreement contained “will assign” language, separate agreement with private company had “hereby assign” language, and Bayh-Dole Act did not change the ordinary rules of employee invention assignment contracts); Advanced Video Techs. LLC v. HTC Corp., 879 F.3d 1314, 1317-18 (Fed. Cir 2018) (invention assignment agreement failed where it stated only that employee “will assign” rights, which is not a present conveyance of future conceptions); St. Clair Intellectual Prop. Consultants, Inc. v. Palm, Inc., No. 06-404-JJF-LPS, 2009 U.S. Dist. LEXIS 37512, *34-35 (D. Del. May 4, 2009) (finding that language in invention assignment agreements did not create a present conveyance of future rights); Freedom Wireless, Inc. v. Boston Comm. Grp., Inc., 220 F. Supp. 2d 16, 19 (D. Mass. 2002) (“In order for a pre-invention assignment contract to create a present assignment of an expectant interest in an invention that automatically vests by operation of law into an actual assignment upon conception, the contract must contain words of present conveyance and must require ‘no further act once an invention [comes] into being.’” (citation omitted) (contrasting contracts that use phrases like “does hereby grant” and “hereby does assign” with a contract that required the inventor to disclose the invention and perform acts necessary to establish ownership); Imatec, Ltd. v. Apple Comput., Inc., 81 F. Supp. 2d 471, 482 (S.D.N.Y. 2000) (“‘I agree to assign and hereby do assign’ . . . constituted a present assignment . . . of future inventions.”); Arachnid, Inc. v. Merit Indus., Inc., 939 F.2d 1574, 1580 (Fed. Cir. 1991) (consulting agreement with phrase “will be assigned” “does not rise to the level of a present assignment of an existing invention, effective to transfer all legal and equitable rights therein”); Filmtec Corp. v. Allied-Signal, Inc., 939 F.2d 1568, 1573 (Fed. Cir. 1991) (invention assignment contract that used the phrase “agrees to grant and does hereby grant” was not merely an agreement to grant future rights, but expressly granted “rights in any future invention”); Treu v. Garrett Corp., 70 Cal. Rptr. 284, 287 (Ct. App. 1968) (where inventor had an assignment clause with his employer, the inventor’s ownership “was a fleeting thing which he had bargained away even before it became a reality . . . the instant the improvement was invented it became the exclusive property of [employer] . . . [plaintiff] had and could have no interest in the invention.”).

[27] See, e.g., Preston v. Marathon Oil Co., 277 P.3d 81, 87-88 (Wyo. 2012) (invention assignment agreement valid when entered into after at-will employment began); Eaton Corp. v. Giere, 971 F.2d 136, 140 (8th Cir. 1992) (both earlier and later invention assignment contracts were valid); Mosser Indus., Inc. v. Hagar, 200 U.S.P.Q. (BNA) 608, 610, 616 (Pa. Ct. Com. Pl. Jan. 11, 1978) (employee signed second invention assignment contract for one dollar); Grove v. Grove Valve & Regulator Co., 84 Cal. Rptr. 300, 303 (Ct. App. 1970) (employee’s valid invention assignment contract was “backdated” by more than a year).

[28] See Iconix, Inc. v. Tokuda, 457 F. Supp. 2d 969, 989, 992 (N.D. Cal. 2006).

[29] By analogy, a federal court in Ohio declined to find that a group of departing employees violated their duties of loyalty by making preparations to compete before leaving where, among other things, the employer’s evidence that they had used its resources was limited to a de minimis claim that one of them spoke to an insurer about the planned new venture “on his business cell phone.” See Fitness Experience, Inc. v. TFC Fitness Equip, Inc., 355 F. Supp. 2d 877, 892-93 (N.D. Ohio 2004) (employees who planned new company and met with attorneys, insurers, and realtors granted summary judgment). The same result might apply for similar de minimis infractions in the invention assignment context.

[30] E.g., Cal. Lab. Code § 2870 (West 1979).

[31] See Cadence Design Sys., Inc. v. Bhandari, No. C 07-00823 MHP, 2007 U.S. Dist. LEXIS 83078, at *21-23 (N.D. Cal. Nov. 8, 2007) (explaining scope of invention assignment in detail and rejecting argument that scope of assignment extends only to specific department in which the employee works).

[32] An attorney had even warned the employee on this point. See id. at *21-22.

[33] While there is little case law on the question, courts have rejected employees’ assertions that company managers or other employees waived or otherwise acquiesced to their personal use of inventions created during employment on two grounds: (a) because an invention automatically vests in the employer at the moment of creation, before the time of any claimed waiver; and/or (b) because the employee did not completely or fully disclose every facet of the invention to the employer, the employer could not have waived an interest in something it did not know everything about. See, e.g., DDB Techs., L.L.C. v. MLB Advanced Media, L.P., 517 F.3d 1284, 1289-90 (Fed. Cir. 2008) (employee’s arguments of waiver and estoppel based on manager’s disinterest in idea were irrelevant because contract gave rise to automatic assignment at moment of creation); Iconix, 457 F. Supp. 2d at 986-87 (employee’s claim that manager and general counsel had expressed disinterest in the idea did not prevent invention assignment where idea was never fully disclosed). See also Eaton Corp., 971 F.2d at 137-38 (employee told employer just before his resignation “of his plans to develop and market his own competitive device” but invention vested in employer under invention assignment contract).

[34] See Applied Materials, Inc. v. Advanced Micro-Fabrication Equip. (Shanghai) Co., 630 F. Supp. 2d 1084, 1089-1091 (N.D. Cal. 2009) (voiding a clause under California law that created a purportedly rebuttable presumption that the employer owned inventions for one year after the employee left, even where no trade secrets were misappropriated); Ingersoll-Rand Co. v. Ciavatta, 542 A.2d 879, 890-95 (N.J. 1988) (applying a reasonableness analysis to a post-employment holdover invention assignment clause; finding that such covenants might be enforceable beyond “trade secrets and confidential information” to include some additional, poorly-defined category of “highly specialized, current information not generally known in the industry” that appears indistinguishable from contemporary definitions of what trade secret law covers); Fed. Screw Works v. Interface Sys., Inc., 569 F. Supp. 1562, 1564 (E.D. Mich. 1983) (agreement which required employees to assign all inventions for an indefinite time period relating to the employer’s field of activity or “contemplated field of activity” found unenforceable); Armorlite Lens Co. v. Campbell, 340 F. Supp. 273, 275 (S.D. Cal. 1972) (finding clause in employee invention assignment agreement which required assignment of new ideas and concepts for one year after employment ended was unenforceable unless such inventions used the trade secrets of the former employer); Winston Research Corp. v. Minn. Mining & Mfg. Co., 350 F.2d 134, 145-46 (9th Cir. 1965) (where employer’s contract required invention assignment of post-employment inventions that were “based upon” its confidential information, court affirmed rulings where such inventions were premised on former employer’s trade secrets, and thus owned by employer, or where no trade secrets were used); De Long Corp. v. Lucas, 176 F. Supp. 104, 127-28 (S.D.N.Y. 1959) (where settlement agreement between former employer and former employee only assigned rights to inventions developed during employment, and where evidence was that new inventions post-dated employment, former employee owed no assignment obligation), aff’d, 278 F.2d 804 (2d Cir. 1960); Guth v. Minn. Mining & Mfg. Co., 72 F.2d 385, 388 (7th Cir. 1934) (invention assignment agreement without time limitation was “contrary to public policy”); cf. Yield Dynamics, Inc. v. TEA Sys. Corp., 66 Cal. Rptr. 3d. 1, 26 (Ct. App. 2007) (where employer’s invention assignment contract created a one-year, post-employment holdover stating that there was a rebuttable presumption that inventions conceived within one year of leaving are owned by the employer, and that inventions during that time must be disclosed, court did not address legality of the clause and affirmed trial court on grounds that employee disclosed a post-employment development to the former employer, which did not request assignment).

[35] See Milliken & Co. v. Morin, 731 S.E.2d 288, 295 (S.C. 2012) (finding a one-year holdover clause enforceable even though it covered inventions created post-employment and without use of the employer’s trade secrets).

[36] For a well-written student note critiquing holdover clauses, see Marc B. Hershovitz, Note, Unhitching the Trailer Clause: The Rights of Inventive Employees and Their Employers, 3 J. Intell. Prop. L. 187, 206-12 (1995).

[37] See Feeney v. Transition Automation, Inc., No. 06-11677, 2008 WL 190766, at *17-19 (D. Mass. Jan. 9, 2008) (where individual disclosed invention before becoming employee and signing invention assignment contract, court interpreted language of contract to vest ownership with individual, not employer); Voith Hydro, Inc. v. Hydro W. Group, Inc., No. C–96–1170 SC, 1997 WL 154400, at *6 (N.D. Cal. Mar. 26, 1997) (although company had a valid employee invention assignment with individual, company did not own invention that employee had conceived and disclosed before becoming an employee of the company); Bailey v. Chattem, Inc., 684 F.2d 386, 391 (6th Cir. 1982) (employer did not own employee’s idea that “was complete prior to the time he began to work for” the employer, and thus prior to invention assignment agreement); Fox v. Kingsland, 81 F. Supp. 433, 437-38 (D.D.C. 1948) (where inventor was employed by Patent Office and also the Signal Corps, he retained ownership despite written agreement with the Signal Corps because invention pre-dated that employment).

[38] The incorporation of pre-existing material does not defeat an invention assignment of the new combination as a whole. See, e.g., ViChip Corp. v. Lee, 438 F. Supp. 2d 1087, 1095 (N.D. Cal. 2006) (granting summary judgment and rejecting employee’s argument that some of the technology in the invention was “preexisting”; employee failed to identify such material or explain how it could be “carved out from” the overall invention that was subject to his employment agreement); Mosser Indus., Inc. v. Hagar, 200 U.S.P.Q. (BNA) 608, 613-14 (Pa. Ct. Com. Pl. Jan. 11, 1978) (granting injunction for employer; even if employee thought up “individual elements” of his creation before he joined the employer, his “combination of those elements into a working prototype occurred during” his employment, and thus the invention assignment clause applied to that “combination”).

[39] See, e.g., Imatec, Ltd. v. Apple Comput., Inc., 81 F. Supp. 2d 471, 482 (S.D.N.Y. 2000) (prior employer had invention assignment contract which covered invention, and thus former employee had nothing to assign to third party and third party lacked standing to sue over patent at issue); Filmtec Corp. v. Allied-Signal, Inc., 939 F.2d 1568, 1573-74 (Fed. Cir. 1991) (stating same rule and remanding for factual findings as to whether company had actual or “inquiry” notice of a prior invention assignment to a former employer or a third party and thus may not have been bona fide purchaser); Thompson v. Automatic Fire Prot. Co., 211 F. 120, 121 (2d Cir. 1914) (where inventor already had assigned invention to plaintiff, and where defendants knew of that assignment, assignment to defendants was invalid: “It seems to us that a business man of reasonable care and prudence would, under these circumstances, before putting his money into an enterprise, have gone to [plaintiff] and asked him if he was making any claim to this invention of [inventor] and, if he said he was, would have asked him what was the nature of his claim, so that the inquirer might advise himself whether he could safely purchase.”).

[40] See Venture Corp. v. Barrett, No. 5:13-cv-03384-PSG, 2015 U.S. Dist. LEXIS 165809, at *13-14 (N.D. Cal. Dec. 9, 2015) (affirming jury finding that employee’s invention belonged to employer because it was created using the employer’s time or resources); Blackbird Techs., Inc. v. Joshi, No. 5:15-cv-04272-EJD, 2015 U.S. Dist. LEXIS 136505, at *24 n.6 (N.D. Cal. Oct. 6, 2015) (issuing preliminary injunction under duty of loyalty theory against employee who started competitive business while employed; terms of injunction barred use of inventions covered by invention assignment agreement); Preston v Marathon Oil Co., 684 F.3d 1276, 1359-60 (Fed. Cir. 2012) (invention assignment agreement valid and assigned rights in invention to employer); NovelAire Techs., LLC v. Harrison, 50 So. 3d 913, 919-21 (La. Ct. App. 2010) (affirming finding that employee violated invention assignment agreement by failing to disclose invention he created and took to new venture after departing); Iconix, Inc. v. Tokuda, 457 F. Supp. 2d 969, 989-92 (N.D. Cal. 2006) (issuing preliminary injunction over social media tool that related to company’s advertising strategy plan); Regents of the Univ. of N.M. v. Knight, 321 F.3d 1111, 1118 (Fed. Cir. 2003) (affirming finding that university patent policy vested ownership in invention in institution); Fenn v. Yale Univ., 283 F. Supp. 2d 615, 629-30 (D. Conn. 2003) (academic bound by university’s patent assignment policies, despite policy amendments since he agreed to it, and thus university owned patent he had licensed to a third party); Univ. of W. Va. v. Van Voorhies, 278 F.3d 1288, 1298 (Fed. Cir. 2002) (academic bound to assign patent under university’s patent policy); Waterjet Tech., Inc. v. Flow Int’l Corp., 996 P.2d 598, 601-02 (Wash. 2000) (where former employee refused to assign rights over patent to workplace invention, court found notice to employee satisfied Washington’s invention assignment statute such that invention assignment contract was enforceable to vest ownership in employer); Vt. Microsys., Inc. v. Autodesk, Inc., 88 F.3d 142, 150-51 (2d Cir. 1996) (affirming ruling that defendant, who started as intern and later became employee and signed an invention assignment agreement, had no interest in software created for the company); Eaton Corp. v. Giere, 971 F.2d 136, 140 (8th Cir. 1992) (granting summary judgment in favor of employer where employee had breached invention assignment agreement by creating a product within employer’s “actual or demonstrably anticipated research and development,” even though employee’s invention related to a market which employer had “not yet been able to crack”); Cubic Corp. v. Marty, 229 Cal. Rptr. 828, 835 (Ct. App. 1986) (affirming judgment for employer where assignment provision was for “inventions which relate to the ‘actual’ or ‘demonstrably anticipated’ business of the employer); Goldwasser v. Smith Corona Corp., 817 F. Supp. 263, 276 (D. Conn. 1993) (entering summary judgment for IBM, noting that employee’s invention or idea which “involve[d] the entry of text into a computer relates to IBM’s ‘actual or anticipated business’”); Gen. Signal Corp. v. Primary Flow Signal, Inc., No. 85-0471B, 1987 WL 147798, at *4 (D.R.I. July 27, 1987) (granting employer ownership of invention where employee was subject to invention assignment agreement, and claimed to have come up with new ideas just five days after leaving job; court found employee’s account “difficult to believe”); Syntex Ophthalmics, Inc. v. Tsuetaki, 701 F.2d 677, 682 (7th Cir. 1983) (affirming finding in favor of employer on invention assignment dispute given “the inherent improbability” of employee’s “story that he conceived of the [invention] four days after he left” his job); Andreaggi v. Relis, 408 A.2d 455, 460 (N.J. Super. Ct. Ch. Div. 1979) (invention related to employee’s work and thus fell within invention assignment agreement); Goodyear Tire & Rubber Co. v. Miller, 22 F.2d 353, 354 (9th Cir. 1927) (where defendant signed an employment contract in which he agreed expressly to assign title to any inventions that related to “methods, processes, or apparatus concerned with the production of any character of goods or materials sold or used” or “any character of goods or materials sold or used” by Goodyear, employer owned invention that fell within general range of what employee was hired to do).

[41] See Mattel, Inc. v. MGA Entm’t, Inc., 616 F.3d 904, 909-12 (9th Cir 2010) (vacating constructive trust for former employer because invention assignment agreement did not contain the word “ideas” and remanding for consideration of extrinsic evidence; court notably did not construe California Labor Code section 2870 except for a passing reference in a footnote). One student commentator found fault with the ruling, but erroneously conflated California’s invention assignment statute with the work for hire doctrine. See Connor Boyd, Note, The Bratz Trap: Ownership and Infringement at the Nexus of Copyright and Employment Law, 45 U.C. Davis L. Rev. 221, 237, 245 (2011) (asserting that section 2870 “refined” the work for hire doctrine, assuming that a state statute could alter the federal Copyright Act and that its coverage was not instead for trade secrets and patentable inventions).

[42] One court denied (in whole or in part) cross-motions for summary judgment where there were disputed facts regarding an employee’s development of source code on his own time, for an unrelated side business, but that the employer contended contained “similarities” with its own source code. See Enreach Tech., Inc. v. Embedded Internet Sols., Inc., 403 F. Supp. 2d 968, 974-75 (N.D. Cal. 2005). See also Rothschild v. Cree, Inc., 711 F. Supp. 2d 173, 182-83 (D. Mass. 2010) (finding fact issues on jurisdictional matters as to individual’s standing to sue for patent infringement where it was unclear whether individual had developed key patent ideas while under another company’s invention assignment agreement or afterwards).

[43] See cases cited supra notes 7, 34, 36. As with any attempt to count holdings from reported cases, caveats are in order. The LEXIS and Westlaw databases do not pick up every ruling in every case, especially those from state trial courts. Some rulings in trade secret cases may be under seal, so that details are not viewable to the public. Appellate rulings matter more than trial court rulings, but the degree of careful analysis varies from case to case. Reported cases may or may not reflect typical fact patterns. Many disputes settle without litigation, or settle before a court issues a ruling on issues of substantive law. As a result, apparent trends in case law can be somewhat arbitrary, as they depend on what lawsuits become easily available to researchers.

[44] Of the three, one remanded on a rather unique contract interpretation, one is unpublished, and one arose from a time period before California enacted its current invention assignment statute. The same outcome under today’s statute seems unlikely. See Mattel, 616 F.3d at 909-12; Applera Corp. v. Illumina, Inc., 375 Fed. App’x. 12, 17-18 (Fed. Cir. 2010) (in a case decided under an older version of California’s invention assignment statute, the court affirmed a finding that employee-attorney’s invention did not “result from” work for the plaintiff); NeoNetworks, Inc. v. Cree, A07-0729, 2008 Minn. App. Unpub. LEXIS 565, at *15 (Ct. App. May 20, 2008) (affirming dismissal of claims against former employees and costs award in their favor after failed ex-employers causes of action, including breach of contract over invention assignment, were rejected where defendants began project after company failed and did not use company resources).

[45] Cmty. for Creative Non-Violence v. Reid, 490 U.S. 730, 743 (1989)

[46] As the text of the statute indicates, no contract is required for a work for hire to exist. Nonetheless, as with the law of employee invention assignments, companies generally insert work for hire assignment terms into standard-form employment contracts, usually if not always next to the terms regarding invention assignment.

[47] See Reid, 490 U.S. at 743 (key case on the work for hire doctrine); JustMed, Inc. v. Byce, 600 F.3d 1118, 1128 (9th Cir. 2010) (affirming finding that software programmer was employee, giving company ownership of software as work for hire).

[48] Id. at 750-51 (“To determine whether a work is for hire under the Act, a court first should ascertain, using principles of the general common law of agency, whether the work was prepared by an employee or an independent contractor.”).

[49] Whether someone is an employee or an independent contractor for purposes of Copyright Act analysis involves the weighing of numerous factors. Reid lists twelve, none of which are “determinative”: (1) the hiring party’s right to control the manner and means by which the product is accomplished; (2) the skill required; (3) the source of instrumentalities and tools; (4) the location of the work; (5) the duration of the relationship between the parties; (6) whether the hiring party has the right to assign additional projects to the hired party; (7) the extent of the hired party’s discretion over when and how long to work; (8) the method of payment; (9) the hired party’s role in hiring and paying assistants; (10) whether the work is part of the regular business of the hiring party; (11) whether the hiring party is a business; (12) the provision of employee benefits and the tax treatment of the hired party. See id. at 751-52. Some courts have suggested that the tax/benefits classification of the author is the most important factor, or at least a disproportionately important factor. See Hi-Tech Video Prods., Inc. v. Capital Cities/ABC, Inc., 58 F.3d 1093, 1097 (6th Cir. 1995) (factor is a “strong indication of a worker’s employment status”); Kirk v. Harter, 188 F.3d 1005, 1008 (8th Cir. 1999) (finding factor important where hiring party reported author to IRS as an independent contractor); Aymes v. Bonelli, 980 F.2d 857, 862-63 (2d Cir. 1992) (failure to pay taxes and benefits “constitute virtual admissions of [author’s] status by [hiring party] himself;” citing numerous post-Reid cases and saying “every case since Reid that has applied the test has found the hired party to be an independent contractor where the hiring party failed to extend benefits or pay social security taxes.”) (citing cases); Numbers Licensing, LLC v. bVisual USA, Inc., 643 F. Supp, 2d 1245, 1251-52 (E.D. Wash. 2009) (where individual had his own consulting entity and company contracted with that entity, and where entity was responsible for individual’s “payroll obligations, tax obligations, and employee benefits,” those and other factors led to court’s conclusion that individual was independent contractor and software was not a work for hire). One court, however, rejected the “usually dispositive” tax/benefits factor where the author worked full-time, was paid weekly, believed he had a long-term relationship, was shown the hiring party’s source code, worked with the hiring party’s equipment and accessed the hiring party’s servers from home, telecommuted but also worked twice a week at the hiring party’s facility for at least some period, assisted with other assignments, received assistance in software programming from the hiring party, had access to the hiring party’s office and equipment, and met a third party as a representative of the hiring party. See Sasnett v. Convergent Media Sys., Inc., No. 95-12262-NG, 1997 WL 33142149, at *9 (D. Mass. Aug. 29, 1997) (on preliminary injunction request, finding that plaintiffs were likely to succeed on the merits).

Courts have identified other possible factors beyond those set forth in Reid. A possible additional factor could be whether the corporation existed at the time of creation. See Billy-Bob Teeth, Inc. v. Novelty, Inc., 329 F.3d 586, 591 (7th Cir. 2003) (doctrine did not apply where, among other things, “the corporation did not exist when the [works] were authored”; company incorporated the year after creation of works). Another possible additional factor is whether the author worked for others at the same time. See Kirk, 188 F.3d at 1008 (“[T]hroughout his six-year relationship with [hiring party], [author] continued to engage in computer consulting with other companies, a factor suggesting that he was an independent contractor.”); Aymes, 980 F.2d at 864 (finding it probative that author “did occasional work for others at the same time” despite working for two years for hiring party); Alcatel USA, Inc. v. Cisco Sys., Inc., 239 F. Supp. 2d 645, 654 (E.D. Tex. 2002) (noting that author “was simultaneously working as an independent contractor for two other companies”). Yet another possible factor is the hiring party’s own perceptions of whether the author was an employee or not. See Hi-Tech Video Prods., Inc., 58 F.3d at 1099 (fact that hiring party did not perceive assistants as employees weighed in favor of finding them to be independent contractors); see also Alcatel USA, 239 F. Supp. 2d at 654 (finding it significant that hiring party had called author an independent contractor in a prior lawsuit before time “when it finally became advantageous for [hiring party] to treat [author] as its employee in order for it to assert ownership of [author’s] creations”).

[50] See Massingill v. Stream, Ltd., No. 3:08-cv-0091-M, 2009 WL 3163549, at *5 (N.D. Tex. Oct. 1, 2009) (company’s motion for summary judgment on ownership of software program denied; fact issues remained where individual started work on software as contractor, later became employee, and whether the work “continued to be developed during his employment with [company] such that it would constitute a ‘work for hire’ under the Copyright Act.”).

[51] 17 U.S.C. § 101 (2012).

[52] As Nimmer puts it, “Under the current Act, even if a work is prepared by an employee, if it is not prepared ‘within the scope of his or her employment,’ it is not regarded as a ‘work made for hire.’ Therefore, an agreement between employer and employee whereby works prepared by the employee that are not prepared within the scope of employment are nonetheless deemed to be ‘works made for hire,’ will not in itself convert such works into the ‘for hire’ category.” 1 Melville B. Nimmer and David Nimmer, Nimmer on Copyright § 5.03[B][1][b][ii] (Matthew Bender, Rev. Ed. 2018).

[53] I have found no empirical study of employment contract terms assigning works made for hire. That said, many samples are publicly available. My own anecdotal experience, based on viewing thousands of employment agreements in Silicon Valley and elsewhere, drafted by a wide variety of law firms, suggests that most recite that works for hire in the scope of employment are assigned – that is, the typical agreement follows the text of the Copyright Act. A very typical sample, pulled from a website which collects such standard terms, reads: “Employee acknowledges that all original works of authorship which are made by him (solely or jointly with others) within the scope of his employment and which are protectable by copyright are works made for hire, pursuant to the United States Copyright Act (17 U.S.C. SS 101 et seq.), and are solely the property of the Company.” See Law Insider, Works Made for Hire Sample Clauses, https://www.lawinsider.com/clause/works-made-for-hire (last visited Oct. 17, 2018).

Two other points about such contracts are worth noting. First, an employer and an employee can negotiate a different structure by contract – such as a royalty-based scheme – though it seems unlikely that a form contract could undercut rights granted by the Copyright Act through the scope of employment test any more than an invention assignment contract could get around state statutory limits on the scope of such agreements. See generally Comput. Data Sys., Inc. v. Kleinberg, 759 F. Supp. 10, 13-14 (D.D.C. 1990) (triable issue of fact on scope of employment because contract stated that if employee developed certain computer software and agreed to transfer it to employer, he would get a royalty to be negotiated at the time).

Second, as with any matter of employment contracts, questions of timing matter. Some courts have taken a position on the question whether, in general, a post-creation writing can give a hiring party work for hire rights. Because the statute’s text seems to indicate that ownership starts from the time of creation, the Seventh Circuit held in 1992 that there cannot be a work for hire based on a written agreement unless the writing predates the creation of the work. See Schiller & Schmidt, Inc. v. Nordisco Corp., 969 F.2d 410, 413 (7th Cir. 1992) (basing its rule on the need to have certainty). By contrast, the Second Circuit rejected this approach and held that it is possible to have a work for hire based on a post-creation written agreement, so long as that agreement evidences a prior agreement before the time of creation. See Playboy Enters., Inc. v. Dumas, 53 F.3d 549, 559 (2d Cir. 1995) (noting that the Nimmer treatise disagreed with Schiller and holding that “the writing requirement . . . can be met by a writing executed after the work is created, if the writing confirms a prior agreement, either explicit or implicit, made before the creation of the work”; remanding on that factual question). See also Compaq Comput. Corp. v. Ergonome Inc., 210 F. Supp. 2d 839, 843-45 (S.D. Tex. 2001) (following Playboy, holding that test requires showing that the parties “understood at the time the [works] were created that the works were made for hire,” and finding test met where written agreement made express reference to work done the year before; written agreement predated author’s subsequent assignment to another company by several years); Kasten v. Jerrytone, No. 02-421, 2004 WL 1857680, at *6 (E.D. La. Aug. 16, 2004) (noting Playboy rule, finding that agreement at issue did not mention previous works, only future works, but separately holding that contract was modified by acceptance of royalties for pre-agreement works).

[54] Restatement (Second) of Agency § 228 (Am. Law Inst. 1958).

[55] Although rarely cited by the courts, the Restatement (Second) of Agency has ten sections regarding “scope of employment.” For the most part, the focus of these sections is on whether the employer will be liable for tortious or illegal acts by an employee. As a result, many of the annotated cases address such matters as joy rides in the company car and assaults by employees, with little value for work for hire disputes. That said, section 229, “Kind of Conduct Within Scope of Employment,” details factors that might assist courts in work for hire disputes, such as “whether or not the act is one commonly done by such servants,” “the time, place and purpose of the act,” and “the similarity in quality of the act done to the act authorized.” Sections 235 and 236, in turn, address the third factor of the employee’s purpose, or intentions. Section 235 states “An act of a servant is not within the scope of employment if it is done with no intention to perform it as a part of or incident to a service on account of which he is employed.” The comment says “It is the state of the servant’s mind which is material.” Section 236 states “Conduct may be within the scope of employment, although done in part to serve the purposes of the servant or a third person.” That is consistent with the case law cited above. Id. §§ 228-237.

[56] See Genzmer v. Pub. Health Tr. of Miami-Dade Cty., 219 F. Supp. 2d 1275, 1280 (S.D. Fla. 2002) (stating rule: “Courts interpret this test as conjunctive. That is, the party attempting to establish that the work was made for hire must establish all three elements.”); Moonstruck Design, LLC v. Metz, No. 02 CIV. 4025(RWS), 2002 WL 1822927, at *3 (S.D.N.Y. 2002) (stating rule); Quinn v. City of Detroit, 988 F. Supp. 1044, 1052 (E.D. Mich. 1997) (same); Roeslin v. District of Columbia, 921 F. Supp. 793, 798 (D.D.C. 1995) (same); City of Newark v. Beasley, 883 F. Supp. 3, 8 (D.N.J. 1995) (same).

[57] See Avtec Sys. v. Peiffer, 21 F.3d 568, 571 (4th Cir. 1994) (noting that if first factor is met, courts will not favor employee just because work was done off-site); McKenna v. Lee, 318 F. Supp. 2d 296, 301 (E.D.N.C. 2002) (same; prisoner did work in his cell, but license plate design was what he was hired to do); Martin v. City of Indianapolis, 982 F. Supp. 625, 634 (S.D. Ind. 1997) (noting that you can’t avoid the work for hire doctrine by doing the work off-site), aff’d, 192 F.3d 608 (7th Cir. 1999); Marshall v. Miles Labs., Inc., 647 F. Supp. 1326, 1331 (N.D. Ind. 1986) (pre-Reid case; article held to be work for hire even though employee did much work away from the employer’s facility because work related to the things plaintiff was hired to do, and because it was based on research done by another employee for employer); Miller v. CP Chems., 808 F. Supp. 1238, 1243 (D.S.C. 1992) (employee worked off-site on software but “[o]n the other hand, the work was performed during the time period in which he was employed by [employer].”).

[58] See, e.g., U.S. Auto Parts Network, Inc. v. Parts Geek, LLC, 692 F.3d 1009, 1018-19 (9th Cir. 2012) (remanding where there was a triable issue of fact as to whether portion of software written when programmer was an employee was within his scope of employment, where among other things he “lived rent-free in a house that doubled” as a company office and did not “strictly” differentiate between work and personal hours, so that his work was “within authorized work hours and space limits.”).

[59] As of May 2018 when this research was completed, these cases are: Genzmer, 219 F. Supp. 2d at 1275; Molinelli- Freytes v. Univ. of P.R., No. 09-1655(DRD), 2012 U.S. Dist. LEXIS 143262 (D.P.R. Sept. 30, 2012); Fleurimond v. N.Y. Univ., 876 F. Supp. 2d 190 (E.D.N.Y. 2012); Le v. City of Wilmington, 736 F. Supp. 2d 842 (D. Del. 2010); Rouse v. Walter & Assocs., LLC, 513 F. Supp. 2d 1041 (S.D. Iowa 2007); McKenna, 318 F. Supp. 2d at 296; Miller, 808 F. Supp. at 1238; Vanderhurst v. Colorado Mountain Coll. Dist., 16 F. Supp. 2d 1297, 1307 (D. Col. 1998); Lewis v. Activision Blizzard, Inc., No. C 12-1096 CW, 2013 U.S. Dist. LEXIS 149784 at *1 (N.D. Cal. Oct. 17, 2013); and Sterpetti v. E-Brands Acquisition, LLC, No. 6:04-cv-1843-Orl-31DAB, 2006 U.S. Dist. LEXIS 21407 at *1 (M.D. Fla. Apr. 20, 2006).

[60] These are: Avtec, 21 F.3d at 568; Moonstruck, 2002 WL 1822927 at *1; TAP Worldwide, LLC v. Becker, No. CV 10-04903 DMG (JCx), 2010 WL 2757354, at *1 (C.D. Cal. July 12, 2010); PFS Dist. Co. v. Raduechel, 332 F. Supp. 2d 1236 (S.D. Iowa 2004), Roeslin, 921 F. Supp. at 793; Quinn, 988 F. Supp. at 1044; Food Lion, Inc. v. Capital Cities/ABC, Inc., 946 F. Supp. 420 (M.D.N.C. 1996); Beasley, 883 F. Supp. at 3; Cannon Grp., Inc. v. Better Bags, Inc., 250 F. Supp. 2d 893, 901 (S.D. Ohio 2003); and Martin, 982 F. Supp. at 625.

[61] See Ryan Vacca, Work Made for Hire – Analyzing the Multifactor Balancing Test, 42 Fla. St. U. L. Rev. 197 (2014) (thorough history and analysis of case law after Reid on employee versus contractor analysis); Llewellyn Joseph Gibbons, Love’s Labor’s Lost: Marry for Love, Copyright Work Made-for-Hire, and Alienate at Your Leisure, 101 Ky. L.J. 113, 180 (2013) (addressing work for hire in the community property context); Jon M. Garon & Elaine D. Ziff, The Work Made for Hire Doctrine Revisited: Startup and Technology Employee and the Use of Contracts in a Hiring Relationship, 12 Minn. J.L. Sci. & Tech. 489, 505-09 (2011) (addressing work for hire agreements and asserting in passing that, as to California’s invention assignment statute, “[t]he two exceptions in section 2870 reasonably approximate the ‘scope of employment’ prong of the work for hire doctrine, as applied to the state law of assignment of employee work product”); Assaf Jacob, Tort Made for Hire – Reconsidering the CCNV Case, 11 Yale J.L. & Tech. 96, 143-144 (2009) (touching on the scope of employment test, but primarily focusing on whether agency law provides the best fit for assessing employee/contractor status); Deborah Tussey, What Ifs and Other Alternative Intellectual Property and Cyberlaw Story: What If Employees Owned Their Copyrights?, 2008 Mich. St. L. Rev. 233, 242 (2008) (“In short, if Congress had not adopted the work for hire doctrine during the early years of corporatization of the copyright industries, it probably would have had to invent an alternative legal structure to handle the licensing of collaborative works in the later years of the copyright terms of those works.”); Vai Io Lo, Employee Inventions and Works for Hire in Japan: A Comparative Study Against the U.S., Chinese, and German Systems, 16 Temp. Int’l & Comp. L.J. 279, 305-06 (2002) (noting, in survey, that Japanese law requires “reasonable compensation if the employee assigns the right to obtain a patent, the patent right, or grant an exclusive license to the employer.”); Carolyn M. Salzmann, Comment, You Commissioned It, You Bought It, But Do You Own It? The Work for Hire: Why Is Something So Simple, So Complicated?, 31 U. Tol. L. Rev. 497 (2000) (focusing on a writing requirement for work for hire in order to clarify ownership); Colleen Creamer Fielkow, Note, Clashing Rights Under United States Copyright Law: Harmonizing an Employer’s Economic Right with the Artist-Employee’s Moral Rights in a Work Made for Hire, 7 DePaul-LCA J. Art & Ent. L. 218, 220 (1997) (reviewing potential conflicts between work for hire and employee-authors’ moral-rights); Shannon M. Nolley, Note, The Work for Hire Doctrine and the Second Circuit’s Decision in Carter v. Helmsley-Spear, 7 DePaul-LCA J. Art. & Ent. L. 103, 128 (1996) (examining the independent contractor vs. employee aspects of work for hire); Jennifer Sutherland Lubinski, Note, The Work for Hire Doctrine under Community for Creative Non-Violence v. Reid: An Artist’s Fair Weather Friend, 46 Cath. U. L. Rev. 119, 125 (1996) (arguing for revamp of Reid test to better benefit artists and incentivize creation of more works); Alan Hyde & Christopher W. Hager, Promoting the Copyright Act’s Creator-Favoring Presumption: “Works Made for Hire” Under Aymes v. Bonelli & Avtec Systems, Inc. v. Peiffer, 71 Denv. U. L. Rev. 693, 716-717 (1994) (primarily examining employee versus contractor case law and approving the concept that absent a written agreement, authors may often own their creations); Anne Marie Hill, Note, The “Work for Hire” Definition in the Copyright Act of 1976: Conflict Over Specially Ordered or Commissioned Works, 74 Cornell L. Rev. 559, 564-66 (1989) (early work assessing pre-Reid case law); Jon L. Roberts, Work Made for Hire: The Fiction, the Reality, and the Impact Upon Software Development, 1 Harv. J. L. & Tech. 97, 128 (1988) (addressing early employee versus contractor case law); Rochelle Cooper Dreyfus, The Creative Employee and the Copyright Act of 1976, 54 U. Chi. L. Rev. 590, 638-47 (1987) (examining work for hire law primarily in the context of academic versus university control over scholarly output).

[62] That treatise provides a general summary, with only a few case citations on “scope of employment” disputes, outside of a long section of disputes between universities and academics. See 1 Melville B. Nimmer and David Nimmer, Nimmer on Copyright §§ 5.03[B][1][b][i]-[ii] (Matthew Bender, Rev. Ed. 2018).

[63] David Loretto, Note, Employee Patents on Computer-Implemented Inventions: The Conundrum of Separate Ownership of Patent and Copyright, 23 Cardozo L. Rev. 705, 711 n.35, 727-28 (2002) (noting possibility of conflicting rights in software covered by copyright and trade secret law and primarily addressing potential conflicts between business-method patents (as patent law stood at that time), employer shop rights, and copyright law). See also Joshua L. Simmons, Inventions Made for Hire, 2 N.Y.U. Intell. Prop. & Ent. L. 1, 46 (2012) (comparing development of work for hire doctrine to incremental development of patent-related concepts like the hired-to-invent doctrine and the shop right doctrine, and arguing that patent law should cease requiring written inventor assignments and instead adopt an umbrella ownership doctrine like work for hire).

[64] Michael D. Birnhack, Who Owns Bratz? The Integration of Copyright and Employment Law, 20 Intell. Prop. Media & Ent. L.J. 95, 152-53, 157-59, 162 (2009) (seeking “the most efficient and fair rule of initial allocation of copyright in works created by authors in the workplace”).

[65] See Catherine L. Fisk, Authors at Work: The Origins of the Work-for-Hire Doctrine, 15 Yale J.L. & Human. 1 (2003) (covering the nineteenth century, the creation of the work for hire doctrine under the 1909 Copyright Act, and modern developments).

[66] For cases where the employer wins on factor one, see Lewis v. Activision Blizzard, Inc., No. C 12-1096 CW, 2013 U.S. Dist. LEXIS 149784, at *10 (N.D. Cal. Oct. 17, 2013) (where employer and former employee clashed over employee’s voice recording used in video games, first factor favored employer where employee handbook listed creation of game content as a job role); Molinelli-Freytes v. Univ. of P.R., No. :09-1655(DRD), 2012 U.S. Dist. LEXIS 143262, at *32-37 (D.P.R. Sept. 30, 2012) (granting summary judgment to employer over faculty member’s proposal for a graduate program; even though employee worked on vacations and holidays, that second factor mattered less where work was clearly of the type employee was hired to perform); Fleurimond v. N.Y. Univ., 876 F. Supp. 2d 190, 201-09 (E.D.N.Y. 2012) (granting summary judgment to employer, which met all three factors as to employee who created graphic design while employed to create such things, where she was expected to perform her job in part at home using her own equipment, and where “no reasonable juror could find that she was not motivated in large part to serve the interests of NYU.”); Rouse v. Walter & Assocs., LLC, 513 F. Supp. 2d 1041, 1056-60 (S.D. Iowa 2007) (applying three-part test; on tangled fact pattern where professors argued that software program created in the course of sponsored university research was not a work for hire, court rejected argument because, among other things, the software development was within the duties they were “hired to perform,” the software “was developed and tested on [university] computers” even if some work was done at home, and professors “were motivated at least in part” to further university research when creating the work); Sterpetti v. E-Brands Acquisition, LLC, No. 6:04-cv-1843-Orl-31DAB, 2006 U.S. Dist. LEXIS 21407, at *20-22 (M.D. Fla. Apr. 20, 2006) (where parties disputed ownership of pasta-making manual, employer prevailed on first factor because it asked the employee to create the manual and provided feedback, showing its creation was part of employee’s job duties); Genzmer v. Pub. Health Tr., 219 F. Supp. 2d 1275, 1281 (S.D. Fla. 2002) (employee hired as a doctor, but in fact his required research “included the drafting of computer programs” and job description was broad enough to cover “the development of the computer program at issue”; in addition, employee’s supervisor gave him “guidance and praise” for the software); Avtec Sys., Inc. v. Peiffer, No. 92-463-A, 1994 WL 791188, at *3-4 (E.D. Va. 1994) (employer won first factor in software case because employee’s job description including writing programs and implementing computer simulation, and because at his employment, employee had generated similar software, and because employer had been able to use the work to “win a contract”), aff’d, No. 94-2364, 1995 WL 541610, at *1 (4th Cir. Sept. 13, 1995); Miller v. CP Chems., Inc., 808 F. Supp. 1238, 1243 (D.S.C. 1992) (employee who wrote software was “not hired primarily for the development of computer programs,” but “the development of the computer programs was at least incidental to his job responsibilities because it was ‘within the ultimate objective of the principal and an act which it is not unlikely that a servant might do.’”) (quoting Restatement (Second) of Agency § 229 cmt. b (Aᴍ. Lᴀᴡ Inst. 1958)).

[67] See Roeslin v. District of Columbia, 921 F. Supp. 793, 798 (D.D.C. 1995) (“It is unfair for the District to now claim that an activity it discouraged – developing the system – was within the scope of plaintiff’s employment.”).

[68] For cases where the employee wins on factor one, see TAP Worldwide, LLC v. Becker, No. CV 10-04903 DMG (JCx), 2010 WL 2757354, at *3-4 (C.D. Cal. July 12, 2010) (export manager of parts manufacturer, which did not develop or sell software, created software to expedite export shipments, usable by any company for a monthly fee; finding in favor of employee when employer moved for preliminary injunction because “[c]reating software was not within the scope of his job and no one else at the company was tasked with anything similar,” employee identified a business need not specific to his employer “but in the business world generally,” and created it “on his own time, not during working hours”; employer did not even learn about software until it terminated the employee); Pavlica v. Behr, 397 F. Supp. 2d 519, 525 (S.D.N.Y. 2005) (finding triable issue of fact for trial where high school teacher created a program to train high school students to conduct scientific research where teacher did so without prompting or direction from school, and where his intent was a model useful for any high school); Moonstruck Design, LLC v. Metz, 02 Civ. 4025 (RWS), 2002 U.S. Dist. LEXIS 14583, at *12-13 (S.D.N.Y. Aug. 12, 2002) (denying motion for preliminary injunction where former employer failed to meet any of the three tests; jewelry designer was not an employee at the time of creation, was hired only in a sales capacity, did not create the design during working hours, and was motivated by desire to honor his wife and other cancer survivors); PFS Dist. Co. v. Raduechel, 332 F. Supp. 2d 1236, 1248 (S.D. Iowa 2004) (employee was truck driver, not a software developer, so this factor favored employee); Martin v. City of Indianapolis, 982 F. Supp. 625, 633-34 (S.D. Ind. 1997) (going a little against the grain of some other cases cited here, finding that employee won on first factor because only a little of his work for employer involved ornamental metal work, but also because he demanded his own fee paid directly to him for artistic work, and his work on sculpture in question differed significantly from his “normal course of work”), aff’d, 192 F.3d 608 (7th Cir. 1999); Roeslin, 921 F. Supp. at 798 (employee hired as labor economist, not software programmer and was discouraged from pursuing project); City of Newark v. Beasley, 883 F. Supp. 3, 8 (D.N.J. 1995) (employee was police officer, but work on classroom training materials to teach young people not to steal cars was unknown to employer, and employee learned nothing from employer that assisted in project, so factor favored employee); Quinn v. City of Detroit, 988 F. Supp. 1044, 1051 (E.D. Mich. 1997) (employee was an attorney, not a programmer, employer never requested the work).

[69] See Quinn, 988 F. Supp. at 1051 (software created at home but later installed at work).

[70] See Fleurimond, 876 F. Supp. 2d at 201-09 (granting summary judgment to employer, which met all three factors as to employee who created graphic design while employed to create such things, where she was expected to perform her job in part at home using her own equipment, and where “no reasonable juror could find that she was not motivated in large part to serve the interests of NYU.”); Rouse, 513 F. Supp. 2d at 1056-57 (applying three-part test; on tangled fact pattern where professors argued that software program created in the course of sponsored university research was not a work for hire, court rejected argument because, among other things, the software development was within the duties they were “hired to perform,” the software “was developed and tested on [university] computers” even if some work was done at home, and professors “were motivated at least in part” to further university research when creating the work); Sterpetti, 2006 U.S. Dist. LEXIS 21407, at *27-28 (employee created pasta-making manual offsite, but at employer’s direction, so employer prevailed on second factor); Genzmer, 219 F. Supp. 2d at 1282 (employee was involved in research and thus would not have been at clinic seeing patients; “it follows that he would not have developed the software” at the employer’s facility).

[71] See Moonstruck Design, 2002 U.S. Dist. LEXIS 14583, at *12-13 (S.D.N.Y. Aug. 12, 2002) (denying motion for preliminary injunction where former employer failed to meet any of the three tests; jewelry designer was not an employee at the time of creation, was hired only in a sales capacity in any event, did not create the design during working hours, and was motivated by desire to honor his wife and other cancer survivors); Raduechel, 332 F. Supp. 2d at 1248 (software work done at home with employee’s own equipment and software licensed to other companies; employee did later modify software for use at employer but factor favored employee); Martin, 982 F. Supp. at 634 (off-site and on own time); Roeslin, 921 F. Supp. at 798 (employee created software in 3000 hours of work at home, fact that he later used it at work immaterial); Beasley, 883 F. Supp. at 8 (police officer created anti-car theft training materials solely off-site, and thus factor favored employee); Avtec, No. 94-2364, 1994 WL 791188, at *4-5 (employee created software at home and with own equipment, even though he later used it at work).

[72] See, e.g., Miller v. CP Chems., Inc., 808 F. Supp. 1238, 1243 (D.S.C. 1992) (“On the other hand, the work was performed during the time period in which he was employed by [employer].”).

[73] See Food Lion, Inc. v. Capital Cities/ABC, Inc., 946 F. Supp. 420, 421-22 (M.D.N.C. 1996) (journalists working as undercover employees made secret tapes of employer’s working conditions, employer tried to claim the tapes as works for hire to suppress them; the court rejecting this stratagem as not meeting any “reasonable interpretation of scope of employment”).

[74] See Lewis v. Activision Blizzard, Inc., No. C 12-1096 CW, 2013 U.S. Dist. LEXIS 149784, at *13-14 (N.D. Cal. Oct. 17, 2013) (employer easily prevailed where employee created sound recording at employer’s request, for employer’s purposes, and under employer’s supervision); Quinn, 988 F. Supp. at 1052; Miller, 808 F. Supp. at 1243-44 (employee’s software was “created to simplify [employee’s] job and to eliminate errors,” employer asked employee to develop similar software for other company products, and all software related specifically to a product made by the employer).

[75] See Avtec Sys., Inc. v. Peiffer, 21 F.3d 568, 572 (4th Cir. 1994) (as to the third scope of employment factor, the employer must show that the employee “was at least ‘appreciably’ motivated by a desire to further its corporate goals[.]”) (citing Restatement (Second) of Agency § 226 cmt. b (Am. Law Inst. 1958)) (“The fact that the predominant motive of the servant is to benefit himself or a third person does not prevent the act from being within the scope of employment. If the purpose of serving the master’s business actuates the servant to any appreciable extent, the master is subject to liability[.]”)).

[76] For cases where the employer wins on factor three, see Fleurimond v. N.Y. Univ., 876 F. Supp. 2d 190, 201-09 (E.D.N.Y. 2012) (granting summary judgment to employer, which met all three factors as to employee who created graphic design while employed to create such things, where she was expected to perform her job in part at home using her own equipment, and where “no reasonable juror could find that she was not motivated in large part to serve the interests of NYU.”); Le v. City of Wilmington, 736 F. Supp. 2d 842, 845-51 (D. Del. 2010) (on summary judgment by employer, municipal employee hired as Information Analyst who created an “Instant Ticketing” program for a city department of licensing and inspections to use in lieu of paper tickets did not own the software, even though he wrote the software “essentially exclusively on his own time, on his own computer at home,” because his purpose was “to facilitate the City’s business of issuing tickets,” and did so at the direction of a supervisor, received input from other employees, and used code previously used for city-owned software); Rouse v. Walter & Assocs., LLC, 513 F. Supp. 2d 1041, 1056-57 (S.D. Iowa 2007) (applying three-part test; on tangled fact pattern where professors argued that software program created in the course of sponsored university research was not a work for hire, court rejected argument because, among other things, the software development was within the duties they were “hired to perform,” the software “was developed and tested on [university] computers” even if some work was done at home, and professors “were motivated at least in part” to further university research when creating the work); Sterpetti v. E-Brands Acquisition, LLC, No. 6:04-cv-1843-Orl-31DAB, 2006 U.S. Dist. LEXIS 21407, at *18-19 (M.D. Fla. Apr. 20, 2006) (employee’s ideas about “gaining personal benefit” from creation of pasta-making manual insufficient where he clearly created manual at employer’s behest, showing at least a partial motivation to server the employer’s interests); Genzmer v. Pub. Health Tr., 219 F. Supp. 2d 1275, 1283 (S.D. Fla. 2002) (employee claimed vague personal reasons, but evidence showed that he “tailored” software to “fit [employer’s] needs”).

[77] For cases where the employee wins on factor three, see Moonstruck Design, LLC v. Metz, 02 Civ. 4025 (RWS), 2002 U.S. Dist. LEXIS 14583, at *12-13 (S.D.N.Y. Aug. 12, 2002) (denying motion for preliminary injunction where former employer failed to meet any of the three tests; jewelry designer was not an employee at the time of creation, was hired only in a sales capacity in any event, did not create the design during working hours, and was motivated by desire to honor his wife and other cancer survivors); Avtec, 21 F.3d at 572 (“appreciably” standard); Avtec Sys., Inc. v. Peiffer, No. 92-463-A, 1994 WL 791188, at *6 (E.D. Va. 1994) (applying test, programmer created work off-site and later used it at work, but his conduct was consistent with a belief that he was the owner, and employer’s bonus payment could have been made “as a reward for [employee’s] willingness to utilize his own property for the benefit of his employer”); PFS Dist. Co. v. Raduechel, 332 F. Supp. 2d 1236, 1248-49 (S.D. Iowa 2004) (where employee had previously licensed software to other companies and then modified it to work with employer’s system, not enough to favor employer on third factor); Roeslin v. District of Columbia, 921 F. Supp. 793, 798 (D.D.C. 1995) (court found evidence that employee was primarily creating software for self-motivation and to create job opportunities for himself, so fact that software also benefited his employee immaterial); City of Newark v. Beasley, 883 F. Supp. 3, 9 (D.N.J. 1995) (police officer intended that many cities, not just his own, might purchase his training materials; even though he believed his employer might be one customer for his work, that was not enough).

[78] See Martin v. City of Indianapolis, 982 F. Supp. 625, 634-35 (S.D. Ind. 1997), aff’d, 192 F.3d 608 (7th Cir. 1999).

[79] See id.

[80] See Cannon Grp., Inc. v. Better Bags, Inc., 250 F. Supp. 2d 893, 901 (S.D. Ohio 2003). This case is atypical, however, as the employer was not a party. The party holding the assignment was the cross-plaintiff in an infringement case, and the cross-defendant was trying to undermine the cross-plaintiff’s copyright registration. There was nothing in the case about whether the employee had been permitted to do the work for the third party while employed, and the court appeared to be moving hastily to shut down the cross-defendant’s strained attempt to defeat the cross-plaintiff’s copyright. The employer was not present to argue its own case for a work for hire. In other words, courts may be less exacting when an accused third party infringer tries to invalidate the copyright on some technicality.

[81] Some courts, of course, find the evidence in conflict, and hold that a trial is necessary. E.g., Koenig v. Dowdy, No. 5:15-CV-00347-RN, 2017 U.S. Dist. LEXIS 163850, at *19 (E.D.N.C. Sept. 28, 2017) (denying motion for summary judgment as to copyright ownership where, among other things, third work for hire factor was subject to disputed evidence over an architectural house plan created by an employee/architect); Bell v. Maloney, No. 1:16-cv-01193-RLY-DML, 2017 U.S. Dist. LEXIS 111867, at *4-5 (S.D. Ind. July 19, 2017) (triable issues of fact on all three factors where attorney took photo that was used on law firm’s website and parties disputed ownership).

[82] Federal courts agree that trade secret causes of action are not copyright-preempted. E.g., GlobeRanger Corp. v. Software AG USA, Inc., 836 F.3d 477 (5th Cir. 2016) (collecting cases from other circuits and joining them). Notably, with the May 2016 enactment of the federal Defend Trade Secret Act, no preemption analysis is necessary as to a cause of action for trade secret misappropriation brought under federal law. See generally Defend Trade Secrets Act of 2016, Pub. L. No. 114-153, 18 U.S.C. §§ 1831-1839, 130 Stat. 376.

[83] For example, in the well-known California case, Cubic Corp. v. Marty, 229 Cal. Rptr. 828, (Ct. App. 1986), the employee’s work product in dispute included a “manuscript describing his invention,” which surely would have been a work for hire under the facts, where the employer easily prevailed under its invention assignment contract. Id. at 830.

[84] E.g., Cal. Lab. Code §§ 2870-72 (West 1979). As noted, however, it remains unclear whether an employer could meet these tests where the employee had no notice of related work by a large, far-flung employer.

[85] PFS Dist. Co. v. Raduechel, 332 F. Supp. 2d 1236, 1240 (S.D. Iowa 2004).

[86] See id. at 1240-43, 1246-47, 1251-53.

[87] See id. at 1247-49.

[88] See id. at 1249. The court spent all of five sentences on the invention assignment claim, and found that “[d]ue to the insufficiency of the evidence to establish the database software as a ‘work for hire,’ or otherwise rightfully owned by plaintiffs, the Court declines to award injunctive relief that ordinarily might be warranted by this claim.”

[89] TAP Worldwide, LLC v. Becker, No. CV 10-04903 DMG (JCx), 2010 WL 2757354, at *1 (C.D. Cal. July 12, 2010).

[90] See id. at *1, 3.

[91] See id. at *4-5.

[92] See Cal. Lab. Code § 2870 (West 1979).

[93] This factual summary is taken from the Fourth Circuit’s first appellate ruling in the case. Avtec Sys., Inc. v. Peiffer, 21 F.3d 568, 569-70 (4th Cir. 1993). See also Avtec Sys., Inc. v. Peiffer, 805 F. Supp. 1312, 1315-17 (E.D. Va. 1992).

[94] Avtec, 21 F.3d at 575. This logic – that copyright ownership precludes a conflicting trade secret ownership claim in the same information – formed the basis for a 2002 ruling in the Eastern District of Texas, where a contractor’s ownership in a software program overrode the hiring company’s claim to trade secrets in the same program. See Alcatel USA, Inc. v. Cisco Sys., Inc., 239 F. Supp. 2d 645, 659 (E.D. Tex. 2002). That case did not involve the work for hire doctrine because the author was an independent contractor, but its holding is consistent with the theme of this essay, that copyright ownership in some instances can conflict with dueling claims under other categories of intellectual property law.

[95] See Avtec Sys., Inc. v. Peiffer, No. 94-463-A, 1994 U.S. Dist. LEXIS 16946, at *11-19, 25-26 (E.D. Va. 1994) (although the employer prevailed on the first scope of employment factor, the employee won on the second and third factors; “Peiffer’s creation of the Orbit Program was too little actuated by a purpose to serve the master.”; rejecting employer’s trade secret claim and awarding no damages on breach of fiduciary duty claim), aff’d, No. 94-2364, 1995 WL 541610, at *1 (4th Cir. Sept. 13, 1995).

[96] Cases in this area appear to be exceedingly rare. In one odd Texas case, a court found that an employer’s state-law claims against a former employee (including trade secret misappropriation, conversion, and unjust enrichment) were preempted by the Copyright Act – a conclusion that would not be reached today at least as to a trade secret claim. Thus, a potential clash between the work for hire doctrine and trade secret ownership through an invention assignment agreement was averted. See Butler v. Continental Airlines, Inc., 31 S.W.3d 642, 651-52 (Tex. App. 2000).

[97] See DDB Techs., L.L.C. v. MLB Advanced Media, L.P., 517 F.3d 1284, 1289-90 (Fed. Cir. 2008) for the exception to this point.

[98] For example, the law of fiduciary duty may require that certain employees provide an “opportunity” to employers. The next essay in this series will focus on that “preparations to compete” context.

Analyzing Aalmuhammed v. Lee in the Context of Entertainment Industry Employment

Analyzing <i>Aalmuhammed v. Lee </i>in the Context of Entertainment Industry Employment

By: Jennifer Yamin*

Download a PDF version of this article here

In Aalmuhammed v. Lee, the Ninth Circuit established a test for determining whether an individual contributor to a work may qualify as a joint author. The test identified three main factors: 1) the author must superintend the work by exercising control; 2) the putative co-authors must make objective manifestations of a shared intent to be co-authors; and 3) the audience appeal of the work must turn on both contributions and the share of each in its success cannot be appraised. Applying these factors, the court concluded that authorship rights could not be granted to a film consultant hired to assist in the creation of the film Malcolm X despite his sizable contributions to the final product.


By analyzing the unique interplay between intellectual property rights and entertainment industry employment law, this Note explores the harmful effects of the Aalmuhammed test on employment and unions across all types of entertainment works. The Note argues that the Ninth Circuit’s test hinders, rather than furthers Congress’s explicit constitutional duty to promote the growth of the arts. In doing so, the test establishes a dangerous precedent that is incompatible with the modern operation of the entertainment industry and paradoxically is detrimental to the very people it intends to protect: creators. The Note concludes that the Aalmuhammed test should no longer serve as the standard courts rely on to determine authorship rights and offers various proposals for reform.


 

Introduction


It may take two to tango, but it takes far more than two to make a film, television show, or music video. It is show business, after all. The rolling credits at the end of Denzel Washington’s motion picture Fences or an episode of the hit television show Game of Thrones exemplify the vast number and diverse array of participants involved in the creative process. From makeup artists to background dancers to even the boom mic operator, any set will undoubtedly be filled with various creative contributors. The credits may fail to accurately portray, however, the frequency at which individuals in the entertainment industry step outside of their designated roles. A background dancer may suggest new choreography or a film consultant can rewrite entire scenes, partaking in the collaborative process in non-stereotypical ways and influencing the direction or even the outcome of the work. This fluidity has freed the entertainment industry from having to stick to the script, allowing the creative process to thrive in an unbound, collaborative environment.

While such flexibility has fueled Hollywood for decades, it has also led many entertainment industry participants to litigate claims over the authorship rights to various creations. Recent litigation has explored this exact issue: who is an author of the work for the purposes of copyright law? In this capacity, the law of intellectual property plays a formidable role in the entertainment industry because of its ability to either grant or deny authorship status to creators. Aalmuhammed v. Lee[1] exemplifies the ambiguity surrounding which contributors can walk away with authorship rights. In this case, a film consultant that presented evidence that he independently wrote at least two entire scenes, translated Arabic into English for subtitles, and participated in editing tried to gain copyright to the motion picture, but was unable to do so because he was not found to be a joint author to the work.[2] In coming to this decision, the Ninth Circuit articulated a three-factor test to determine whether an individual qualifies as a joint author to a work.[3] However, as this note will show, this test has proven to be inconsistently applied throughout the courts, leading both to confusion and a lack of direction when providing guidelines for determining authorship.

In this Note, I argue that the joint authorship test established by the Ninth Circuit is problematic, particularly with respect to the entertainment industry. Part I addresses the current copyright law landscape regarding authorship following Aalmuhammed. Part II showcases the recent inconsistent and ambiguous applications of the Aalmuhammed test, specifically of the audience appeal factor, as applied to different types of entertainment content, including screenplays, songs, and music videos. Part III introduces the interplay between intellectual property rights and entertainment industry employment law. Part IV discusses how the Aalmuhammed test is inconsistent with the modern operation of the entertainment industry. It argues that the application of the test yields negative effects on entertainment industry employment and unions. Part V offers resolutions to this problem by discussing various proposals for reform. This Note concludes that the Ninth Circuit’s Aalmuhammed test is not only incompatible with the functions of the entertainment industry, but also detrimental to employment in the industry, further demonstrating the need for reform.

I. Copyright Landscape

A. 1976 Copyright Act

The United States Constitution promulgates the Copyright Clause under Article I, Section 8, Clause 8. The Copyright Clause specifies that one of the powers of Congress is “ to promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.”[4] Notably, included within this delegation of power is the specification that Congress is empowered to provide copyright protection to “ authors.”[5]

The prevailing statute for copyright law in the United States is the 1976 Copyright Act, (“ The Act”).[6] Oddly, the words “ author” and “ authorship” have yet to be defined, despite their presence in the Copyright Clause and importance to the statutory scheme. Joint works, however, are defined as “ a work prepared by two or more authors with the intention that their contributions be merged into inseparable or interdependent parts of a unitary whole.”[7] Because “ authorship” is left undefined, the legislative history surrounding the definition and inclusion of “ joint work” is often used to shed light on the congressional intent behind the term. The House Report describes that the “ touchstone” of the joint work question “ is the intention, at the time the writing is done, that the parts be absorbed or combined into an integrated unit, although the parts themselves may be either ‘inseparable’ (as the case of a novel or painting) or ‘interdependent’ (as in the case of a motion picture, opera, or the words and music of a song).”[8]

The Act lists the different types of works of which authors may claim authorship. Works of authorship include the following categories: literary works, musical works (including any accompanying words), dramatic works (including any accompanying music), pantomimes and choreographic works, pictorial, graphic and sculptural works, motion pictures and other audiovisual works, sound recordings, and architectural works.[9] Under section 101, each of these categories is individually defined.[10] Motion pictures, for example, are defined as “ audiovisual works consisting of a series of related images which, when shown in succession, impart an impression of motion, together with accompanying sounds, if any.”[11]

B. Joint Authorship in the Ninth Circuit: Aalmuhammed v. Lee

Beginning in 1991, Warner Brothers along with Spike Lee and his production companies, entered into a contract to make the film Malcom X, based on the book The Autobiography of Malcom X.[12] Lee served as the director, co-writer, and co-producer of the film. The film’s leading actor, Denzel Washington, played Malcom X. Washington sought out Jefri Aalmuhammed to assist with the role specifically because Aalmuhammed had previously made a film about Malcom X, and was known to be well-informed and knowledgeable about the leader’s life.[13]

Although Aalmuhammed did not have a contract with Warner Brothers or the director, he still wanted to be credited as the co-writer of the film. In his complaint, he cited evidence showing that his involvement with the film was substantive.[14] Aalmuhammed’s evidence revealed he had reviewed the shooting script, suggested extensive script revisions, directed Washington and other actors, “ created at least two entire scenes with new characters, translated Arabic into English subtitles, supplied his own voice for voice-overs, selected the proper prayers and religious practices for the characters, and edited parts of the movie during post production.”[15]

Aalmuhammed subsequently sought a copyright with the U.S. Copyright Office. The Office issued him a “ Certificate of Registration,” but noted there were conflicting previous registrations of the film. On November 17, 1995, he filed suit against Spike Lee, Lee’s production companies, and Warner Brothers alleging a variety of claims: breach of implied contract, quantum meruit, unjust enrichment, and federal and state unfair competition claims. Lee sought declaratory relief and an accounting under the Copyright Act.[16] Most of these claims were dismissed under a motion to dismiss or on summary judgment.[17] In February 2000, the Ninth Circuit heard argument in the case. The court heard the copyright claim, the quantum meruit claim, and the unfair competition claim.[18]

The court first addressed Aalmuhammed’s joint authorship claim. To determine whether the motion picture qualified as a joint work, the court looked to the definition and language of “ joint work” in the Act. It concluded that for a work to qualify as a joint work it must: 1) be a copyrightable work; 2) by two or more “ authors;” 3) the authors must intend their contributions to be merged into inseparable or interdependent parts of a unitary whole; and 4) each author is required “ to make an independently copyrightable contribution” to the disputed work.[19]

The court found that Aalmuhammed had established a genuine issue of fact for three of the four elements of a “ joint work.” Specifically, the court noted that Malcolm X was a copyrightable work and intended to be a unitary whole. Aalmuhammed’s evidence that he “ rewrote several specific passages of dialogue that appeared in the film” and “ wrote scenes relating to Malcolm X’s Hajj pilgrimage that were enacted in the movie” was credited by the court as a copyrightable contribution.[20] Further, all parties involved had the intent that Aalmuhammed’s contributions would be merged into interdependent parts of the movie as a unitary whole.[21] Despite the fact that Aalmuhammed readily met these standards, the court ultimately held he did not qualify as an “ author” for the purposes of the Act, and thus found that Aalmuhammed was not one of at least two authors required to establish a “ joint work.”[22]

The court devoted a significant portion of its decision to the joint authorship question. After noting that “ authorship is not the same thing as making a valuable and copyrightable contribution,” the court analyzed the traditional uses and applications of the word “ author,” concluding that the word “ author,” as used in this context, refers to the “ activity of one person sitting at a desk with a pen and writing something for publication.”[23] The idea and definition of “ author,” the Ninth Circuit notes, is “ relatively easy to apply to a novel” and “ to two people who work together in a fairly traditional pen-and-ink way.”[24] The relatively easy application ends, however, when “ the number of contributors grows and the work itself becomes less the product of one or two individuals who create it without much help.”[25] Reiterating the traditional bases of the word, the court added, “ the word is traditionally used to mean the originator or the person who causes something to come into being, or even the first cause, as when Chaucer refers to the ‘Author of Nature.’”[26] Lastly, the court cited the Gilbert and Sullivan song, “ I Am the Very Model of a Modern Major General,” to show that because Gilbert’s words and Sullivan’s tune are inseparable, the audience can know that both were the creative originators or authors.[27] The court then listed several theories that could help establish the authorship of a film. For example, they noted the author may be “ the producer who raises the money.”[28] Or, under Eisenstein’s theory and the “ auteur” theory, the author may be the editor or the director, respectively.[29] Lastly, they stated that “ traditionally, by analogy to books, the author was regarded as the person who writes the screenplay, but often a movie reflects the work of many screenwriters.”[30]

Turning to case law to shed further light on the discussion of who an author may be, the court cited the Supreme Court’s decision in Burrow-Giles Lithographic Co. v. Sarony.[31] In Burrow-Giles, the Court expanded the concept of authorship to include a photographer who exercised a sufficient degree of control over the subject of his photograph. The Court’s support for this decision included various English authorities and the Founding Fathers of the Constitution.[32] Ultimately, the Ninth Circuit concluded that an “ author” is “ the person to whom the work owes its origin and who superintended the whole work, the ‘master mind.’”[33] Applying this definition to a film, the Aalmuhammed court finds that an author is limited “ to someone at the top of the screen credits, sometimes the producer, sometimes the director, possibly the star, or the screenwriter – someone who has artistic control.”[34] And, lastly, the Aalmuhammed court relied on Thomson v. Larson,[35] a Second Circuit decision in which it was affirmed that Larson, a playwright, did not intend joint-authorship with Thomson, the dramaturg for the Broadway musical Rent. The absence of both decision-making authority and billing as a co-author led the Second Circuit to find that the work was not joint.[36]

Extracting from Burrow-Giles, Thomson, and the Gilbert and Sullivan example, the Ninth Circuit compiled a list of three factors that would constitute the criteria to establish who should be considered an author for the purposes of a joint work (assuming there is no contract stating otherwise): 1) the author must superintend the work by exercising control; 2) the putative co-authors must make objective manifestations of a shared intent to be co-authors; and 3) the audience appeal of the work must turn on both contributions and the share of each in its success cannot be appraised.[37] With respect to the first factor, the court specified that “ this will likely be a person ‘who has actually formed the picture by putting the persons in position, and arrang[ed] the place where the people are to be – the man who is the effective cause of that,’” or “ the inventive or master mind” who “ creates, or gives effect to the idea.”[38] The second factor requires the putative co-authors make objective manifestations of a shared intent to be co-authors. For example, listing both “ Gilbert and Sullivan” as the authors of the song would satisfy this requirement. Objective manifestation—as opposed to subjective manifestation—is required because a fraudulent outcome might result if co-authors were to conceal their true intentions regarding a work’s authorship.[39] For example, an author could communicate to the other “ author” that she intended to be co-authors to the work, only to register for a copyright in her own individual name. The court expresses concern over mutual intent under a subjective intent standard as becoming “ an instrument of fraud.”[40] And under the third factor, the audience appeal of the work must turn on both contributions and “ the share of each in its success cannot be appraised.”[41] Notably, the court specified that control will be the most important factor in many cases. They also qualified that the factors “ cannot be reduced to a rigid formula, because the creative relationships to which they apply vary too much.”[42]

In applying these factors, the court first found that Aalmuhammed did not superintend the work, and instead, Warner Brothers and Lee controlled it. Comparing Aalmuhammed to the dramaturg in Thomson, they found that Aalmuhammed may have made useful recommendations, but the film may not have benefited from them unless Lee chose to accept them.[43] Secondly, they concluded that there were no objective manifestations of intent to be co-authors among any of the parties involved.[44] The court did not address the third factor.

The court grounded its decision to withhold authorship from Aalmuhammed in policy concerns. According to the court, granting authorship to the plaintiff in this scenario would frustrate the constitutional goal to “ promote the progress of Science and useful Arts.” The court feared that the threat of losing sole ownership of the work itself may incentivize authors to insulate themselves throughout the creative process.[45] As a result, the overall decrease in collaboration could thereby impoverish the arts in terms of both quantity and quality.

Moreover, adopting Aalmuhammed’s broader definition of joint authorship would allow many other contributors to deprive the sole authors of proper title. Specifically, “ research assistants, editors, and former spouses, lovers and friends would endanger authors who talked with people about what they were doing, if creative copyrightable contribution were all that authorship required.”[46] From the court’s point of view, the fear of being stripped of absolute title thus becomes even more prevalent, and consequently, the urge to resist collaboration as well.

II. Recent Inconsistent and Ambiguous Applications of the Audience Appeal Factor in Entertainment

The judicial system is no stranger to the issue of joint authorship in the entertainment industry. In applying the Aalmuhammed test, however, lower courts have reached inconsistent results when considering various kinds of entertainment works, such as screenplays,[47] songs,[48] and music videos.[49] In dealing with frequent litigation surrounding the joint authorship issue, courts have turned to the Ninth Circuit’s Aalmuhammed decision for guidance. The third factor of the Aalmuhammed joint work analysis, the audience appeal factor, poses particular issues with respect to entertainment works in that it is applied inconsistently and ambiguously to screenplays, songs, and music videos.[50]

A. Screenplays

In Richlin v. MGM, for example, the same court that crafted the audience appeal factor admitted that “ it is nearly impossible to determine how much of [a] Motion Picture’s audience appeal and success can be attributed to the [Screenplay].”[51] The Ninth Circuit suggested that a film’s success might be attributed to a number of factors other than the screenplay itself, such as the actor’s “ legendary comedic performance,” the composer’s “ memorable score,” or the director’s “ award-winning direction.”[52] Despite acknowledging that there were various creative contributions, none of which could have independently and solely drawn in the audience, the court still found that the audience appeal factor favored the screenwriter plaintiffs seeking copyright of the film, because without their original screenplay, the motion picture would not exist. Ultimately, the analysis surrounding the audience appeal factor proved moot, as the court found that the first two factors weighed heavily in favor of the defendant, Metro-Goldwyn-Mayer Pictures, Inc., thereby outweighing the third factor entirely.[53]

The treatment of the third factor in Richlin is problematic in three ways. First, many motion pictures stem from or are a derivative of a screenplay or book. Without the original writing, the motion picture, in many instances, would cease to exist. The conclusion that the audience appeal and success of a work can be attributed to the screenplay simply because the motion picture was derived from it represents an analytical gap in that it forecloses the recognition of any other valuable creative contributions. By limiting its reasoning to this logic, the Ninth Circuit set a standard in which the audience appeal factor of the Aalmuhammed joint work analysis may favor the party who contributed to the original work and seeks rights as a joint author. This approach circumvents the third factor analysis entirely, where the court would ordinarily need to compartmentalize the various creative contributions, such as those of the writers, actors, composers, and directors, to determine which are responsible for the majority of the audience appeal.

Second, had the court performed a proper audience appeal analysis, they would have been met with a burdensome task that could result in divergent results given the subjective interpretations of motion pictures and what factors contribute to their success. The Ninth Circuit, acknowledging that this analysis would be “ nearly impossible,” instead decided that concluding one could not exist without the other is sufficient.[54]

Third, Richlin represents a case in which the first two factors weighed in favor of the defendants’ claim that the parties were not joint authors. The court left no guidance for a case in which the first two factors weigh in favor of the plaintiff seeking to establish joint authorship. In such a situation, it remains unclear if the third factor would be determinative or continue to carry little or no weight at all. The court also failed to explain whether a more thorough analysis, like the one the Ninth Circuit avoided, should take place. And although the Ninth Circuit qualified in Aalmuhammed that the factors could not be boiled down to an exact formula,[55] the court’s reasoning in Richlin leaves a great deal of ambiguity for litigants and courts to confront in future disputes.

B. Songs

There have been inconsistent ways of analyzing the audience appeal factor in the context of songs as well. For example, in Eli Brown, III v. Michael C. Flowers,[56] the plaintiff and defendant had formed a partnership called Hectic Records. Flowers and Brown recorded hundreds of demo R&B songs together, where Brown, in his role as sound engineer and producer, “ added riffs and beats,” thus “ establishing tempo, ambiance, echo, reverberation, treble, bass, frequency, gain, bandwidth, distortion, and equalizing.”[57] Flowers, a songwriter, then traveled to New York and New Jersey to market the songs he and Brown recorded, including “ I Wish,” a track that Brown had engineered and produced. The song was picked up by singer Carl Thomas, who re-recorded the song and released it on an album that eventually went platinum. The partnership later dissolved, and, according to Brown, Flowers “ applied for and received copyright registrations for these recordings” and “ subsequently licensed these recordings to other artists, who, in turn, made re-recordings.”[58] Brown then filed suit against Flowers under the Copyright Act of 1976, alleging that the recordings were joint works and he thus had copyright in them as well.[59] The Fourth Circuit ultimately affirmed the district court’s dismissal of Brown’s federal copyright claim, finding he did not meet the requisite author requirement.[60]

In an unpublished concurrence and dissent in part, Judge Gregory wrote a separate opinion in which he applied the Aalmuhammed three-factor test and reached a result that differed from the majority. He found that “ because Brown allege[d] that he served as the engineer and producer of the master recordings, [he] conclude[d] that [Brown] had significant decision making authority over the substance and form of the master recordings.”[61] Second, he found that Brown had sufficiently alleged that the parties made objective manifestations of their intent to be co-authors because of Brown’s allegations “ that the parties collaborated over a period of four years in recording demo R&B songs under their Hectic Records partnership label and in attempting to distribute those recordings to third parties under that label.”[62] Lastly, the third factor was met “ because the audience appeal of R&B songs is attributable in large measure to their underlying riffs and beats, which, invariably, are incorporated into those songs by recording engineers and producers.”[63]

Although he found joint authorship for Brown, Judge Gregory extracted from the ambiguity surrounding the Aalmuhammed audience appeal factor in ways that other courts may not. For example, despite attempting to apply an objective approach to the analysis, Judge Gregory generalized and promulgated stereotypes about the specific type of work at issue. In doing so, his analysis ran counter to the purpose of the audience appeal factor. His conclusion that the appeal of R&B songs is attributable in large measure to the songs’ “ underlying riffs and beats,” neglects the subjective analysis owed to creative works of music. For example, the song at issue could depart from the expectation or standard that R&B songs are successful because of those specific features. Judge Gregory does not constitute nor represent the “ audience” and their views of the song. The vocalist on the song, for example, could be the sole reason the audience is drawn to the piece. Thus, while Judge Gregory did recognize Brown’s right, he did so in a way that did not provide for sufficient flexibility in the analysis that such creative works of entertainment are owed.

In Ford v. Ray,[64] a claim for joint authorship was dismissed based on the Aalmuhammed test. In this case, Ford, the plaintiff, collaborated with Ray, the defendant, on an album where Ford “ independently created a handful of drum tracks, called ‘beats,’ that he thought would make a good foundation for hip hop songs” as well as “ provided ‘scratching’ for the chorus of the song and a solo.”[65] Ray’s role was to choose which of the beats would be used for Sir-Mix-a-Lot’s Grammy award-winning song, “ Baby Got Back.” Ray, without Ford’s knowledge, filed copyright registrations for some of the songs they worked on together, identifying himself as the sole author.[66] Ford subsequently filed suit alleging the works were “ joint works” and that he was the co-author.

In employing the Aalmuhammed three-factor test, the United States District Court for the Western District of Washington found that the first two factors – that the author must superintend the work by exercising control and the putative co-authors must make objective manifestations of a shared intent to be co-authors – weighed in favor of Ray, and thus concluded that Ford was not a co-author.[67] In acknowledging the audience appeal factor, the court stated that Ford “ fail[ed] to allege facts from which one could conclude that his contributions account for the appeal of ‘Baby Got Back.’”[68] The court sought more information such as “ how much of the music was [Ford’s] creation and how much was edited, programmed, and altered by [Ray].”[69] They found the fact that Ray used one of Ford’s “ drum tracks ‘as the basis for the song’ and incorporated [Ford’s] scratching” insufficient.[70] Relying on the complaint, the court concluded that Ford had made no contribution to the lyrics nor did he have any control over the music after handing it over to Ray.[71] In coming to this conclusion, the court acknowledged that the audience appeal factor remains “ something of an unknown” to the analysis and granted the dismissal based solely on the first two factors.[72]

Brown and Ford are thus paradigmatic of how the Aalmuhammed test, particularly the audience appeal factor, has been interpreted, analyzed, and implemented inconsistently when applied to songs.

C. Music Videos

Litigation surrounding music video joint authorship issues further exemplifies the ambiguous application of the audience appeal factor with respect to various kinds of entertainment works. In Morrill v. The Smashing Pumpkins,[73] Morrill was a member of the band The Marked. The band produced a music video titled “ Video Marked” which had been shown at various clubs where The Marked performed in order to promote their group.[74] Defendant Corgan, one of the members of The Marked and the sole permanent member of The Smashing Pumpkins, was a member of The Marked at the time the music video was made, but later left the group. Years later, defendants Corgan, The Smashing Pumpkins, and Virgin Records America released a video titled “ Vieuphoria” which featured short clips of images from “ Video Marked.”[75] Morrill brought suit against the defendants, alleging he was the sole owner of the copyright for the music video. The United States District Court for the Central District of California found that the first two factors favored a finding of a joint work.[76] With respect to the audience appeal factor, the court concluded that the audience appeal can be attributed to both parties because the appeal of “ Video Marked” “ was most likely based on the audience’s ability to view images of a younger Corgan. This is suggested by the packaging for ‘Vieuphoria,’ which advertises super secret, super special extra stuff shot by the band.”[77] The court focused on Corgan’s appearance as fulfilling the audience appeal requirement which ostensibly “ rests both on the video’s visual aspects and on the composition and performance of the music.”[78]

In Eagle Rock Entertainment, Inc. v. Coming Home Productions, Inc.,[79] the Central District of California faced an ownership dispute over “ Guns, God and Government,” a live concert video album filmed during shock rock musician Marilyn Manson’s controversial 1999-2000 world tour.[80] Both parties claimed to have produced the video by contributing to the editing and production of it. The court found the parties to be joint authors by analyzing the audience appeal factor and relying on Morrill. Unlike in Morrill, however, the court did not specifically address which aspects contributed to the visual appeal or the composition and performance of the music. Thus, the court seemed to employ the same standard to measure audience appeal differently in that it specifically sought out and emphasized an extra element in one music video, but did not require or address it in the other.

III. The Interplay Between Intellectual Property Rights and Entertainment Industry Employment Law

The Aalmuhammed test proves significant even outside of the courtroom, where its application has a direct impact on the livelihoods of those who participate in the entertainment industry. Recent examples in pop culture reflect the close link between intellectual property rights and employment. Take Steven Jan Vander Ark, for example, the Harry Potter fanatic who sought to publish an encyclopedia[81] featuring the many characters, settings, and overall magic of J.K. Rowling’s popular book series.[82] Rowling and Warner Brothers Entertainment sued Vander Ark seeking to enjoin the publication of the book alleging it infringed their copyrights given the substantially similar, if not equivalent, nature of the content.[83] While Rowling walked away victorious, Vander Ark was left sobbing on the stand stating that Rowling’s public denunciation of his book “ had ostracized him from the ‘Harry Potter community.’”[84] Why was Rowling, a prominent, successful author of a world-famous series, so keen on stopping one author—or more properly labeled, a fan—from producing a Harry Potter encyclopedia? And why was Vander Ark left so concerned about his reputation among other Harry Potter fans?

Intellectual property rights hold so much force in the realm of entertainment employment that even renowned authors like J.K. Rowling are adamant about protecting them. As of November 2018, the U.S. media and entertainment industry is a $735 billion market, representing a third of the global industry.[85] In 2017, the arts, entertainment, and recreation industry employed 2,370,160 people.[86] Duncan Crabtree-Ireland notes that “ [a]lmost all workers in the entertainment industry . . . are dependent on intellectual property law (and particularly copyright) for their livelihoods.”[87] Overall, the interplay between intellectual property and employment is a unique one in entertainment, where intellectual property rights serve as a policing mechanism to ensure that parties are rewarded for original works of authorship by establishing standards and thresholds for artists to qualify for the benefits of employment.

Despite the significant effect of intellectual property rights on employment opportunities, practices within the entertainment industry often make it difficult to determine the appropriate allocation of rights. First, the industry is inherently filled with intangible content that without proper regulation could be subject to taking.[88] Furthermore, technological developments have augmented the means through which audiences consume content, and facilitated the dissemination of entertainment content throughout the consumer marketplace. Cynthia Baron explains this phenomenon, noting how the “ increasing convergence of film and new media ‘has simultaneously increased the availability of film and turned it into part of a data stream where images become information that is simply passing through.’”[89] In this capacity, “ film has also become part of the flow,” blurring and stretching the origins of the content farther away from the original artist or author. [90] As a result, it becomes a challenge to pinpoint the original source or author of a work, leaving the claimant to intellectual property rights with a daunting task. Justice Scalia voiced similar concerns in Dastar Corporation v. Twentieth Century Fox Film Corporation,[91] stating “ without a copyrighted work as the basepoint, the word ‘origin’ has no discernable limits” and “ figuring out who is in the line of ‘origin’” would prove difficult.[92] The high transaction costs for identifying and crediting sources may serve to decrease, or even paralyze, creative cultural production.

Second, the length and pace of projects in entertainment production make it difficult to monitor the contributions and rights of those who work on entertainment projects. Employment in Hollywood’s film industry, in particular, is project-based so that employers can “ quickly assemble personnel with highly specialized skills for a short period of time. Producers have no incentive to offer long-term contracts because informationally complex jobs are difficult to monitor.”[93] In hiring for short-term projects:

[T]he entertainment industries exist on ideas turned into deals. When an idea is ‘hot,’ immediate action is desired. Parties rush to agree, and, in the process, desire at times outraces common sense. The ‘deal,’ as it turns out, is strictly verbal, or there are scattered memos but no single, final, formal written agreement. If the production proceeds as envisioned . . . [t]here is no problem because the idea becomes a deal that produces a success, and everyone is happy.[94]

As a result of the quick pace of deal-making, oral contracts have replaced written agreements as the primary means of memorializing deals. The industry reliance on oral contracts—often referred to as the “ handshake deal”[95]—reflects the priority of completing projects on time. This priority, in turn, comes at the expense of fleshing out the full details of an employment agreement, which can suppress workers’ efforts to obtain credit for their contribution. If two parties orally and mutually agree to certain terms of a relationship or a course of conduct, the agreement becomes enforceable regardless of whether or not a written agreement was signed.[96] Main Line Pictures, Inc. v. Basinger,[97] one of the pivotal cases involving oral contracts in Hollywood, epitomizes the ubiquity and enforceability of oral contracts in the film industry.

In 1990, Main Line Pictures asked actress Kim Basinger to star in the movie “ Boxing Helena.” On several occasions, Basinger orally agreed to star in the film. A year and a half later, Main Line learned through rumors that Basinger no longer intended to act in the film as she had agreed to do. Main Line subsequently filed a complaint alleging breach of an oral and written contract.[98] The jury concluded that Basinger had breached her oral contract with Main Line and awarded them damages.[99] On appeal, the California Court of Appeals took note of and gave credence to the practice of oral contracts in entertainment, explaining “ because timing is critical, film industry contracts are frequently oral agreements based on unsigned ‘deal memos.’”[100] Even though this case was later reversed on a technicality and settled, it reiterates the control and influence of oral contracts in the industry.

Another common practice unique to the film industry is what Mihaela Mihailova has called “ invisible labor,” where many contributors or collaborators to a film often go unrecognized.[101] She notes that “ Hollywood labor history is marred by a host of such obfuscations,”[102] pointing to unrecognized contributions in Darren Aronofsky’s film Black Swan as exemplifying such an omission of credit. In Black Swan, actress Natalie Portman was largely recognized as the star of the film, displaying professional ballet movements on screen. In reality, however, ballet dancer Sarah Lane was the one who performed most of the challenging dance scenes. With the help of post-production technology, Lane’s face was simply replaced with Portman’s. Neither Lane nor the animation team were credited in the marketing or promotional materials, and instead Portman walked away with the accolades, winning an Academy Award.[103] Mihailova sums up this practice as a “ policy of deliberately suppressing any public acknowledgment of the invisible labor that facilitates the star’s performance.”[104]

Entertainment industry employees are heavily reliant on the recognition of intellectual property rights for employment opportunities. Industry practices, however, make it difficult to determine who can claim those rights. As will be discussed further, authorship can directly affect the salaries, royalties, and overall benefits associated with recognition. In addition to the numerous barriers these participants have to overcome, such as the implications of intangible property, the dominance of oral contracts, or the “ invisible labor” theory, the Aalmuhammed test only exacerbates the difficulties creators in entertainment must face.

IV. Why the Aalmuhammed Test is Problematic for the Entertainment Industry

A. The Aalmuhammed test is incompatible with the operation of the modern Entertainment Industry

The creative process is anything but streamlined and Hollywood’s productions are no exception. While certain structures and processes exist as guidelines, deviation occurs far more often than in other industries. Many factors inherent to the filmmaking process lead to situations in which contributors take on roles outside of their designated responsibilities, such that even in the case of a formal written contract, departures from official roles are frequently made. Thus, while the credits to a film may describe the job titles that participants in the creative process held, they should by no means convey the message that the producer, for example, stuck solely to the responsibilities with which a producer is conventionally associated. The overall creative process behind filmmaking involves many people who played a role, small or large, in the creation of the film.

The three-factor test articulated in Aalmuhammed to determine joint authorship, however, does not adhere to the highly collaborative nature of the filmmaking process, nor to the widespread custom of creative collaboration in entertainment in general. One of the underlying reasons the test is incompatible with the contemporary practices of entertainment is that it is extracted from antiquated sources and thus has proven incongruous in the modern industry. The Ninth Circuit heavily rooted its analysis of “ author” in tradition, citing cases from 1884, Eisenstein, Chaucer, and the Founding Fathers.[105] Instead, the Court should have drawn from these sources with greater generality to maintain an appropriate balance between tradition and innovation. For example, the court could have taken the idea of the writer as the “ person behind the pen”[106] and applied it more liberally to the context at issue. While Aalmuhammed may not have been the physical person holding the pen and writing the screenplay, he did write entire scenes as well as much of the dialogue that was incorporated into the final product.

Scholars also support the argument that the Aalmuhammed test diverges from the Act’s original intent and, as a result, is inconsistent with the needs and goals of the entertainment industry. Professors Shyamrkishna Balganesh, Justin Hughes, Peter Menell, and David Nimmer argue that “ interpreting the statute so narrowly ignores the careful study and negotiation that went into crafting the Section 101 definitions.”[107] The professors assert that the definition of “ joint work” included in Section 101 “ recognizes a wide range of collaborative working arrangements by requiring only that the authors collaborated with each other, or if each of the authors prepared his or her contribution with the knowledge and intention that it would be merged with the contributions of other authors.”[108] Further, the use of the mastermind standard could be interpreted as an evasive strategy in that it could be used for the sake of “ administrative convenience or to avoid unjustified windfalls, not out of fidelity to legislative intent.”[109] They further note:

Such an interpretation misses the broad and open-ended recognition of collaborative creativity that Congress intended. Undue emphasis on singular control (“ the . . . mastermind”) is antithetical to the very nature of joint authorship, which is an intrinsically collaborative exercise. The nature of a collaborative enterprise is such that at times different authors will exercise more control than the others over the work. To require a contributor to exercise equal “ inventive” control in order to be a joint author is therefore unrealistic. Second, under the control standard, it is impossible for contributing authors to know in advance whether they are exercising sufficient control over the unitary work while making their individual contributions. The joint authorship doctrine thereby becomes unpredictable, defeating the “ paramount goal” of the 1976 Act.[110]

The Ninth Circuit also mistakenly assumed that the factors apply neutrally to all subject matter. The reliance on the Gilbert and Sullivan example assumes that a songwriter and film consultant have made equal contributions throughout the creation of the work and to the final product in identical ways. The creative process behind the song that Gilbert and Sullivan wrote, however, is extremely simplified and not representative of the creation of a motion picture.[111] Similarly, the court relied on Thomson and assumed that the process of creating a Broadway musical mimics that of a motion picture.[112] Another assumption undergirding the Ninth Circuit’s opinion is that the creative process for one motion picture is the same for all other motion pictures.[113] The function and operation of creative practices within the industry are far from uniform, and to assume so was a dangerous judicial mistake.

The test is also incompatible with the “ above the line” versus “ below the line” distinction in entertainment, which plays a primary role in monitoring the types of relationships formed during the creative process.[114] Those in Hollywood who are “ above the line” are recognized as the creative talent or financial supporters of a project.[115] The people “ below the line,” on the other hand, hold the technical and behind-the-scenes positions.[116] In particular, the audience appeal factor of the test clashes with the first factor because the audience appeal of a film, more often than not, is owed to the “ above the line” personnel such as the A-list stars like Meryl Streep or Emma Stone, who “ would not qualify under the court’s ‘inventive or master mind’ test.”[117] Instead, the director of the film is usually credited and recognized as the mastermind of the film. If these famous participants are unable to secure copyright interests in a work under such a test, it is further unlikely that a person who is “ below the line” and part of the “ invisible labor” that goes into a film, would either. Despite the frequency with which the Aalmuhammed test is employed by numerous courts when discussing various entertainment works, the test has proven inapplicable to certain categories within the modern entertainment industry.

B. The Aalmuhammed test negatively affects entertainment industry employment

The inability of the Aalmuhammed test to adapt to the entertainment industry’s practices yields unintended consequences on employment. Generally, joint authorship is an important intellectual property right that may play a part in establishing whether or not an individual can obtain future employment opportunities. If an individual meets the requirements to qualify as a joint author, then she will be able to register and obtain a copyright in the work. This further enables her to receive future benefits associated with the copyright, such as the opportunity to be involved with any sequel or a work tangentially related to the original work. This will then increase her exposure, and therefore her potential to garner greater recognition or reputation in the industry. Being known as a joint author may also signify to other creators her potential as a collaborator for future projects, thereby opening the door to future employment opportunities. Without recognition of joint authorship, a co-author may be deprived of the employment opportunities arising from or incidental to copyright ownership.

There are several more specific ways in which the Aalmuhammed test is unable to best serve the interests of those seeking employment and those currently employed in the entertainment industry. First, the audience appeal factor disincentivizes participants from contributing to a work because it does not adhere to the fluidity of the creative process. This is evidenced in Ford, where the plaintiff did not provide enough evidence upfront that his contributions throughout the creative process impacted the audience appeal of the work. Creators may therefore take additional precautions external to the creative process itself in order to abide by the types of inquiries a court may make upon reviewing whether an artist contributed enough to qualify for authorship rights. The creative process is a lengthy one, involving a broad spectrum of contribution ranging from the first spark of creativity to the final product. Drafts, shot suggestions, lighting changes, deletions, and many other steps that blur the process frequently occur. Requiring or expecting parties involved in the music creative process, for example, to memorialize agreements and document material information at all stages of production is a large burden, especially given the prevalence of oral contracts in the entertainment industry. The difficulty of proving specific contributions after the fact arguably discourages collaboration between creators, and, as a consequence, may adversely affect the labor market within the industry as well.

Second, the Aalmuhammed test reinforces the monopolization of the creative process evident in Hollywood and in doing so, inhibits the opportunities for employment growth in entertainment. A recent deal between the Walt Disney Company and 21st Century Fox, in which the former acquired the latter’s assets for $52.4 billion epitomizes how the “ Hollywood heavyweights” continue to maintain and exert the upper hand in the broader entertainment industry.[118] By expanding its empire even further, Disney undoubtedly will gain a strong foothold over 21st Century Fox’s operations, employees included, and even “ more control over some of the content that fuels [Disney’s] business.”[119]

Even for smaller scale projects in entertainment, employers normally will be able to fulfill positions needed for a music video or a film because prospective creators are eager for their “ big break” in Hollywood and would not normally pass down a shot at fame. Those with established reputations such as Steven Spielberg and Quentin Tarantino, then, are able to maintain the upper hand in terms of bargaining power (and the subsequent copyrights), making it much more difficult for individuals to make a name for themselves. Different hiring structures or novel approaches to the distribution of responsibility will not be implemented as a result of the cemented status quo and monopoly over employment. Opportunities for employment growth and diversification are thus decreased in ways that harm those trying to break into the industry.

Third, as a consequence of increased litigation over co-authorship claims, courts have experimented with different approaches to applying the Aalmuhammed standard, resulting in inconsistent results and ambiguity for jobseekers and employers alike. As a result, proper fair notice is not offered to those who would benefit from awareness of the requirements of co-authorship under the Aalmuhammed test. This lack of prospectivity with respect to authorship rights will have a detrimental effect on entertainment industry employment because those who are hiring will not know what protocol or processes to put in place regarding the allocation of degree of control and distribution of responsibilities. That courts may or may not accord proper weight to the audience appeal factor leaves entertainment industry employers with little guidance for dealing with the co-authorship issue and avoiding potential lawsuits over joint authorship claims.

Similarly, creators or authors will lack the necessary guidance for understanding whether or not they have to direct their contributions in a way that will affect the audience appeal of the work. One major incentive a participant to a project may have in agreeing to provide services is the potential recognition he or she could gain as a result of being listed as a co-author to a work. The incentive may be lost among the confusion caused by the unclear rules and judicially constructed obstacles blocking the path to legal recognition of joint authorship. Both parties involved may then lose the incentives needed for a successful employment relationship, thus decreasing the potential for collaborative creation.

Fourth, a false impression has been created that any standard other than the Aalmuhammed test will result in the overbroad dissemination of copyrights. This, in turn, makes those with more bargaining power overprotected while they employ others. In Garcia v. Google, for example, the Ninth Circuit relied on Aalmuhammed when addressing Google’s argument that an actress hired to say a few lines in a film posted on YouTube did not qualify as a joint author to the work.[120] The court cited Aalmuhammed for the proposition that a work cannot be defined based on “ some minimal level of creativity or originality” because such a definition “ would be too broad and indeterminate to be useful.”[121] Further, the court noted that “ its animating concern” in Aalmuhammed was that such a definition of “ ‘work’ would fragment copyright protection for the unitary film . . . into many little pieces” where anyone might qualify as an author.[122] The Google court feared this would make “ Swiss cheese of copyrights.”[123] The court similarly stated that “ treating every acting performance as an independent work would not only be a logistical and financial nightmare, it would turn cast of thousands into a new mantra: copyright of thousands.”[124] The Ninth Circuit took a risk-averse approach, fearing that any alternative to the Aalmuhammed test may open the floodgates to copyright claims of joint authorship by every cast and crew member that worked on a film. By establishing Aalmuhammed as the only alternative to a world in which joint authorship claims run amok, the desire to keep Aalmuhammed as the standard is increased, as are concurrently, the limits on employment.

C. The Aalmuhammed test has a negative effect on entertainment unions, which, in turn, further harms industry workers’ employment opportunities

The Aalmuhammed test may also yield detrimental effects on employment by indirectly diluting the efficacy of entertainment unions to adequately represent workers. Specifically, the inability of the test to properly identify rights may lead to union strikes which, as has happened in the past, could cause networks to restructure their employment processes such that opportunities for post-production creation will be limited.

One of the unions in the entertainment industry that may be negatively affected by this phenomenon is the Writers Guild of America West (WGAW). The WGAW “ is a labor union composed of the thousands of writers who write the television shows, movies, news programs, documentaries, animation, videogames and new media content that keeps audiences constantly entertained and informed.”[125] The Guild covers the rights of writers with respect to derivative works, specifically “ minimums, residuals, credits, pension and health contributions, separated rights, and more.”[126] Strikes are an effective union bargaining tactic. Normally, union members strike in an effort to convince or coerce employers to change their employment practices and conditions. Common wisdom may suggest that the potential of the Aalmuhammed test to strip authors of the rights and recognition they deserve might encourage members of the WGAW, for example, to strike. However, striking in response to being denied recognition rights can harm union members’ prospects for future employment.

The 2007 Writers Guild of America (WGA) strike against ABC Studios exemplifies the potential negative effects of a labor strike on the writers. Writers for the hit television show Lost went on strike in order to receive better recognition and compensation for their creations of derivative digital content tied to the original show.[127] During this time, innovations in technology had led to the proliferation of derivative works, including “ blogs, alternate reality games (ARGs), and mobisodes[128] to promote the series.”[129] Recognizing the rights issues that arose from the growth of these markets, the WGA increased its support for enterprising authors or writers who sought to challenge the traditional studio employment model.[130] For example, the WGA website stated, “ [i]n the age of YouTube, Hulu, Crackle, and MyDamnChannel, new media outlets and digital technologies provide writers increased opportunities to become true creative entrepreneurs, armed with the tools and distribution channels necessary to connect directly with audiences – and often without studio/network intervention – like never before.”[131] While the WGA was increasing its support for such “ entrepreneurial” writers, Denise Mann points out that following the strike:

[R]ather than embrace the mass collaborative approach to television explored during the Lost moment, the ‘big three’ networks appear to have bolstered their traditional bureaucratic fortresses to maintain singular control over all aspects of the broadcast business. In particular, ABC fired several of the principal executives involved in the Lost franchise and reabsorbed many of the functions that the Lost writers had performed in collaboration with the network’s programming, marketing, licensing, and merchandising divisions. Since 2011, the big three have reverted to more conventional programming choices (reality shows, sit-coms, episodic dramas), reasserted control over their licensed properties (computer games, novels, board games, and the like), and expanded their in-house digital marketing divisions to create and manage digital promotions tied to their series — all, it seems, in an effort to maintain stricter controls over their industry in the post-strike environment.[132]

Mann further observes that, since the strike, studios and networks have turned to low-cost in-house labor in what appears to be an effort to escape the demands of writers seeking greater recognition and remuneration for their contributions.[133] She concludes that powerful networks like ABC have displayed an “ unwillingness to collaborate with creative partners to the degree seen during the Lost moment.”[134] More troublesome, however, is that “ all of the networks have shown a reluctance to grant additional power to showrunners who wish to engage fans by embracing social media as a viable component of the television experience in the digital age.”[135] She describes that the networks are hesitant about working with outside partners, such as “ WGA-represented transmedia czars, PGA-represented transmedia producers, super-fans who wish to contribute to the television experience, or the type of creative entrepreneurs that YouTube is hiring for its 100-channel partnerships.”[136] Overall, Mann noticed a trend of studios and networks trying to minimize the threat of collaboration by “ taking a giant step backward toward their analog past.”[137]

The treatment of the producers for the hit television show Heroes further demonstrates how networks have responded to the post-production creative boom in unfavorable ways. In 2007, the show’s producers proposed to network executives the formation of a “ transmedia team” to be tasked with managing the continuity of any extensions of the show’s narrative as it expanded beyond its traditional television medium and format through the creation of bonus material made available on other media platforms. The responsibilities of this team would be divided into three main roles: “ merchandising,” “ coordinat[ing] all narrative mobilizations of the property across comics, the Internet, and the like,” and “ work[ing] with the stars of the series to secure their participation in promotions and content made for these new media.”[138] However, the network ultimately dismissed the proposed model for maintaining narrative continuity and fidelity. This is a paradigm example of what Derek Johnson classifies as the “ contradiction between singular authorship and the decentralized creativity of networked production cultures.”[139] Regrettably, it appears that participants who seek alternative means of employment are only met by rejection.

Operating as an obstacle to the recognition of judicial rights and providing limited guidance on creators’ legal rights, the Aalmuhammed test arguably contributes to a growing dissatisfaction among entertainment industry unions. The outcome has been debilitating in certain instances. Although some minimums or residuals may be established in the short-term, networks may put up more barriers in response to such proposed changes in the long-term, as exemplified by the Lost strike and the doomed Heroes proposal. Given the weaker bargaining position of union members, big networks are well-placed to structure their operations in a manner that promotes their interests, much to the detriment of entertainment creators. Union mobilization would be disrupted and potentially paralyzed if networks take advantage of the Aalmuhammed test protections. Employee creators would struggle to avail themselves of joint authorship status under the three-factor Aalmuhammed test, especially in the emerging context of derivative supplemental and promotional content designed to “ keep the story alive” as it migrates across consumption platforms. Creators would be hard-pressed to show that they contributed to the original story, that all relevant parties manifested mutual intent to be co-authors, and that they contributed to audience appeal.

Writers are also directly burdened by the adverse effects the Aalmuhammed test has on entertainment unions. Given the monopolization of the creative process, writers may struggle to obtain a writing credit for projects they worked on despite their sizable contributions. The WGA has a system in which “ any writer whose work represents a contribution of more than 33% of a screenplay shall be entitled to screenplay credit. One exception exists for original screenplays in which any subsequent writer or writing team must contribute 50% to the final screenplay.”[140] In order to determine whether a writer meets this threshold, an arbiter is brought in to consider various elements of their contribution to the work at issue, including dramatic construction, original and different scenes, characterization or character relationships, and dialogue.[141] Each work is looked at independently and there are no formulas an arbiter must use. The WGA, in contrast to the Aalmuhammed test, approaches the authorship analysis with much needed flexibility.

In Hollywood, receiving a writing credit signifies the writer’s accomplishment and can bolster her reputation. A writing credit is an important touchstone of the writer’s professional development because as “ writers move from project to project, a career is created as people move from credit to credit.”[142] Further, a writing credit “ facilitates the assessment of talent in a high-velocity labor market” and provides “ residuals . . . [to] . . . compensate writers during periods of slack employment, thus keeping their human capital in the industry.”[143] One way in which a residual provides better compensation to writers is through the allocation of a credit bonus, which is “ a provision of an individual hiring contract stating that the writer will receive a bonus if the writer is determined to get screen credit.”[144] By minimizing the writer’s ability to legally establish co-authorship, the Aalmuhammed test may have adverse economic repercussions on writers by minimizing their ability to gain writing credit and its associated benefits.

Overall, the Aalmuhammed test-or-nothing approach creates obstacles to joint authorship status and future employment opportunities for creative contributors.

V. Proposals for Reform

That brings the analysis to the next question: what should be the standard by which joint authorship is determined?

Professor Balganesh et al. recommend implementing a test that analyzes the contributions made and the “ mutual intent among the collaborators that they be joint authors,” rather than focusing on the control element of the work.[145] Such inquiry as to mutual intent will only arise when the “ parties have not expressly contracted ownership.”[146] This approach would allow “ all principal creative collaborators in the production of a motion picture, sound recording, or other collaborative work” to “ qualify as joint authors”.[147] While the professors qualify that this would not result in every single contributor being granted joint authorship, this standard remains problematic with respect to the entertainment industry. For example, agreements to take on certain roles often do not go as far as establishing a mutual intent to be joint authors given the time constraints of production on set and the prevalence of oral contracts.

Professor Christopher Jon Sprigman proposes another systematic approach that imposes a default rule that parties may contract around to reach the idealized authorship relationship.[148] In practice, a default rule would establish which party would automatically be granted joint authorship status. Sprigman further suggests making the default rule “ painful” so that the parties, in objecting to it, will have to provide information about which contributors they would prefer to credit with joint authorship status.[149] Parties have the ability to opt-out of the default rule and can negotiate against the background of clear rules. While this proposal offers an organized, approachable solution, it may not pair well with the elasticity and flexibility of the creative process. Specifically, there may not be time to engage in the contracting, nor would there be widespread knowledge of a default rule one would need to contract out of in the first place. Further, it places the parties in a situation where a negotiation, or at least a discussion, would have to take place from the outset. This may disincentivize collaboration, as creators may wish to avoid the trouble and cost of bargaining over authorship ex ante.

Another area of the law that has been proffered as a solution to the authorship question is the work made for hire doctrine. The Copyright Act defines a work made for hire as “ a work prepared by an employee within the scope of his or her employment” or a work that is commissioned specifically to be used in a variety of listed works, including, in most relevant part, “ as a part of a motion picture,” with the parties “ expressly agree[ing] in a written instrument signed by them that the work shall be considered a work made for hire.”[150] In Community for Creative Non-Violence v. Reid, the Supreme Court delineated a list of factors that can be used to determine whether or not a work qualifies as a work made for hire by identifying the status of the person hired.[151] Hiring status becomes determinative to the extent that the copyright lies with the employer if the creative contributor is an employee or independent contractor with the requisite signed written agreement. To determine the employment status of a hired party, the Supreme Court looked to the following factors:

the hiring party’s right to control the manner and means by which the product is accomplished, the skill required, the source of the instrumentalities and tools, the location of the work, the duration of the relationship between the parties, whether the hiring party has the right to assign additional projects to the hired party, the extent of the hired party’s discretion over when and how long to work, the method of payment, the hired party’s role in hiring and paying assistants, whether the work is part of the regular business of the hiring party, whether the hiring party is in business, the provision of employee benefits, and the tax treatment of the hired party.[152]

The proposed solution, then, is to have the employers tailor the course of employment to fit the Reid factors such that the person hired would affirmatively qualify as an employee. Employers would arguably use this strategy when they could anticipate a potential copyright claim. As evidenced by the fluctuations inherent to the entertainment industry in both hiring and throughout the creative process itself, predicting copyright claims or even framing the employment relationship to adhere to the Reid factors is, for the most part, too difficult to execute.

While more compatible with the functions and operations of the modern entertainment industry, a case-by-case analysis lacks the guidance courts may seek when addressing such joint authorship issues. Stripping courts of judicial guidance on the one hand, and rote application of an obsolete and poorly designed three-part test on the other, are both not ideal. Instead, a middle-ground approach should be adopted in order to provide the courts with necessary assessment tools and background knowledge, while still allowing them the freedom to embrace the fluidity they may encounter in the various fact patterns presented.

First, guidelines should be tailored to the subject or content at issue, whether it takes the form of films, television shows, musicals, music videos, or the like. Tests with distinct factors may be developed for the various kinds of content, as long as they allow for the flexibility inherent to the entertainment industry and permit creators to freely roam within its carefully crafted boundaries. For example, a test for motion pictures could account for creative contributions by film consultants by analyzing the quantum of control they exercised over the film. Under such a test, a court may determine that if Aalmuhammed had rewritten or had a significant impact on more than one half of the total scenes, then he should have been awarded joint authorship status. Musical works, too, should be analyzed in ways that account for more than stereotypical elements that are considered by a court to make a “ rap” song popular or a “ country” song identifiable. Instead, an objective approach should be implemented whereby the stereotypes that Judge Gregory had relied on in Flowers are removed.

Second, plaintiffs or defendants may introduce as evidence market studies that describe which aspects of an entertainment work served as most attractive to its audience members. By relying on what consumers actually think, this may reduce the potential for mistaken attributions of ownership and disputes over whose contributions made the biggest impact with respect to revenue or audience appeal. Market studies of this kind can also assuage the concerns of judges who fear the theoretical onslaught of endless co-authorship claims. This middle-ground approach will thus avoid any slippery slope, or as Google’s attorneys would like to say, “ the Swiss cheese of copyrights.”[153]

Another player that can improve the current framework surrounding joint authorship is Congress. Congress can amend the Copyright Act in ways that provide guidance not just to judges, but also to the creators themselves. For example, the Section 101 definition of “ joint work,” as well as the Copyright Act in general, currently lack a definition of “ author.” Congress can narrowly revise the Copyright Act to account for idiosyncratic works to which there are many contributors, by qualifying that “ an author, particularly in creative works in which many different collaborators contribute, may not necessarily fit the ideal or traditional characteristics one would normally associate with the title.” Such a revision would clarify overbroad or vague language, while also loosening the standards to which collaborators may have felt previously bound. Such guidance could function as the proper notice collaborators should be able to rely on, thus diminishing any reluctance on the part of would-be collaborators to enter into a collaborative creative relationship, reducing the likelihood of authorship battles erupting, and ultimately improving the overall efficiency of the entertainment industry.

Conclusion

Then he burst out crying. ‘Sorry,’ he said, regaining his composure. ‘It’s been difficult because there’s been a lot of criticism, obviously, and that was never the intention.’”[154] – Steven Jan Vander Ark


The judicial system’s current treatment of those who seek vindication of their claims for joint authorship paints a rather bleak picture: one where enthusiastic creators like Steven Jan Vander Ark are reduced to tears on the stand. Vander Ark is joined by film consultants like Aalmuhammed, screenwriters like Richlin, sound engineers like Brown, and band members like Morrill who put their trust in the legal system to secure their authorship rights as creative contributors after their fellow collaborators denied them the credit and benefits they deserve. To their dismay, the Ninth Circuit’s three-part test to determine whether an individual qualifies as a joint author to a work was unaccommodating with respect to both the major creative roles they played as well as the modern functions of the entertainment industry.

Entertainment has proven to be one of the most booming industries in society today, providing high quality content in unprecedented ways across the globe. Yet while the industry continues to grow, the legal framework for the authorship question remains stuck in the past. Poorly suited for the expanded marketplace for collaboration and unchanged in the face of widespread innovative media technologies, the Ninth Circuit’s obsolete test hinders, rather than furthers Congress’s explicit constitutional duty to promote the growth of the arts. This paradox, where the judicial system’s treatment of authorship ultimately disincentivizes authors, the very people expected to create, heightens the necessity for reform. To improve employment in the modern entertainment industry and further the policy goals of copyright law, the Ninth Circuit’s test must no longer serve as the governing standard for evaluating joint authorship rights.

The judicial system must play its role in granting and reinforcing the allocation of intellectual property rights so that those who are deserving of credit in the creative process can do so without worrying about having their contribution minimized. This is a call to action for the courts to recognize its role in creating barriers to employment in the entertainment industry and craft more appropriately tailored authorship tests to better accommodate those who are employed and who seek to be employed in the entertainment industry. A more carefully articulated framework can be implemented to strike a fair balance between infinite copyrights on the one hand and a monopoly over copyrights on the other. It is not solely up to the courts; Congress must amend the Copyright Act to overcome the inherent vagueness underlying the statutory definition of “ author” as well. These reforms have been advanced by scholars and creators alike. In a multibillion dollar industry where success is highly correlated with authorship rights, it is simply too costly to leave the Vander Arks of the world stripped of their rightful legal protection.


*J.D. Candidate, New York University School of Law, 2019; B.A., Communication Studies, Northwestern University, 2016. The author would like to thank Professor Day Krolik and the JIPEL Notes Program participants: Jared Greenfield, Vincent Honrubia, Chloe Kaufman and James Yang.

[1] Aalmuhammed v. Lee, 202 F.3d 1227 (9th Cir. 2000).

[2] Id. at 1230.

[3] Id. at 1234.

[4] U.S. Const. art. I, § 8, cl. 8.

[5] Id.

[6] See generally, 17 U.S.C. §§ 101 et seq. (2012).

[7] Id. § 101.

[8] H.R. Rep. No. 94-1476, at 120 (1976), reprinted in 1976 U.S.C.C.A.N. 5659, 5736.

[9] 17 U.S.C. § 102.

[10] Id. § 101.

[11] Id.

[12] Aalmuhammed v. Lee, 202 F.3d 1227, 1229 (9th Cir. 2000).

[13] Id.

[14] Id. at 1230.

[15] Id.

[16] Id.

[17] Id.

[18] Id.

[19] Id. at 1231.

[20] Id. at 1231-32.

[21] Id. at 1232.

[22] Id. at 1236.

[23] Id. at 1232.

[24] Id.

[25] Id.

[26] Id.

[27] Id.

[28] Id.

[29] Id.

[30] Id.

[31] Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53, 54 (1884).

[32] Id. at 57-59.

[33] Aalmuhammed, 202 F.3d at 1233.

[34] Id.

[35] Thomson v. Larson, 147 F.3d 195, 207 (2d Cir. 1998).

[36] Id.

[37] Aalmuhammed, 202 F.3d at 1234.

[38] Id.

[39] Id.

[40] Id.

[41] Id.

[42] Id. at 1235.

[43] Id.

[44] Id.

[45] Id.

[46] Id. at 1235-36.

[47] See, e.g., Complaint at 1, Horror, Inc. v. Victor Miller, No. 3:16-cv-01442 (D. Conn. Aug. 24, 2016) (Victor Miller, the author of the screenplay Friday the 13th, issued termination notices to the film’s producers to notify them that he was reclaiming his copyright in the original screenplay); Ron Dicker, Filmmaker Says ‘Stranger Things’ Creators Stole His Ideas in New Lawsuit, Huffington Post (Apr. 4, 2018, 10:39 AM), https://www.huffingtonpost.com/entry/stranger-things-suit-duffer-brothers_us_5ac4b7b7e4b093a1eb20d34c (Charlie Kessler filed suit against Matt and Ross Duffer alleging they stole his ideas for a sci-fi series called “ Montauk” in their popular Netflix series “ Stranger Things”).

[48] See, e.g., Complaint at 4, Brittle v. Time Warner, Inc., No. 3:16-cv-00908-JAG (E.D. Va. Nov. 11, 2016) (Gerald Brittle sued Time Warner and other entertainment company defendants over the exclusive motion picture rights to the film The Conjuring based on the book he wrote called The Demonologist).

[49] Complaint, Frank Ocean v. Om’Mas Keith, No. 2:18-cv-01383 (C.D. Cal. Feb. 20, 2018) (Musical artist Frank Ocean brought a declaratory action of copyright non-ownership against music producer Om’Mas Keith, who claimed that he had co-written musical compositions with Ocean for the singer’s album titled Blonde).

[50] Aalmuhammed, 202 F.3d at 1234. (“ [T]he audience appeal of the work must turn on both contributions and the share of each in its success cannot be appraised.”) (internal quotation marks omitted).

[51] Richlin v. Metro-Goldwyn-Mayer Pictures, Inc., 531 F.3d 962, 970 (9th Cir. 2008).

[52] Id.

[53] Id.

[54] Id.

[55] Aalmuhammed v. Lee, 202 F.3d 1227, 1235 (9th Cir. 2000).

[56] Brown v. Flowers, 196 F. App’x 178, 184 (4th Cir. 2006).

[57] Id.

[58] Id.

[59] Id. at 184.

[60] Id. at 181

[61] Id. at 191 (Gregory, J., concurring in part and dissenting in part).

[62] Id.

[63] Id.

[64] Ford v. Ray, 130 F. Supp. 3d 1358, 1359 (W.D. Wash. 2015).

[65] Id.

[66] Id.

[67] Id. at 1364.

[68] Id. at 1363.

[69] Id. at 1364.

[70] Id. at 1363-64.

[71] Id. at 1364.

[72] Id.

[73] Morrill v. Smashing Pumpkins, 157 F. Supp. 2d 1120 (C.D. Cal. 2001).

[74] Id. at 1121.

[75] Id.

[76] Id. at 1126.

[77] Id. at 1125 (internal quotation marks omitted).

[78] Id.

[79] Eagle Rock Entm’t, Inc. v. Coming Home Prods., Inc., No. CV 03-571 FMC(AJWx), 2004 WL 5642002, at *1 (C.D. Cal. Sept. 1, 2004).

[80] Id. at *3.

[81] See generally, Harry Potter Wish List, http://hpwishlist.warnerbros.com/ (last visited Nov. 11, 2018) (providing an example of an extrinsic encyclopedia and fan-site dedicated to J.K. Rowling’s book series).

[82] John Eligon, Rowling Wins Lawsuit Against Potter Lexicon, N.Y. Times, Sept. 8, 2008, at B3.

[83] Warner Bros. Entm’t Inc. v. RDR Books, 575 F. Supp. 2d 513 (S.D.N.Y. 2008).

[84] Eligon, supra note 82.

[85] Media and Entertainment Spotlight: The Media and Entertainment Industry in the United States, Selectusa, https://www.selectusa.gov/media-entertainment-industry-united-states (last visited Nov. 11, 2018).

[86] May 2017 National Industry-Specific Occupational Employment and Wage Estimates, Sector 71: Arts, Entertainment, and Recreation, U.S. Dep’t of Lab., https://www.bls.gov/oes/current/naics2_71.htm.

[87] Duncan Crabtree-Ireland, Labor Law in the Entertainment Industry Supplemental Payments, Intellectual Property Rights, and the Role of Unions, 31 Ent. & Sports L. 4, 4–5 (2014).

[88] Id.

[89] Cynthia Baron, The Modern Entertainment Marketplace, 2000-Present, in Acting 144 (Claudia Springer & Julie Levinson eds., 2015) (quoting author Stephen Keane).

[90] Id.

[91] Dastar Corp. v. Twentieth Century Fox Film Corp., 539 U.S. 23 (2003).

[92] Id. at 35.

[93] William T. Bielby & Denise D. Bielby, Organizational Mediation of Project-Based Labor Markets: Talent Agencies and the Careers of Screenwriters, 64 Am. Soc. Rev. 64, 66 (1999).

[94] Gary M. McLaughlin, Oral Contracts in the Entertainment Industry, 1 Va. Sports & Ent. L.J. 101, 119 (2001).

[95] Id. at 120.

[96] See Jay M. Spillane, Lawsuits Over "Handshake Deals" Are As Old As the Entertainment Industry (and Can Be Easily Avoided), 11 Ent. & Sports L. 15, 16 (1993).

[97] Main Line Pictures, Inc. v. Basinger, No. B077509, 1994 WL 814244, at *1 (Cal. Ct. App. Sept. 22, 1994).

[98] Id. at *4.

[99] Id. at *5.

[100] Id. at *2.

[101] See Mihaela Mihailova, Collaboration Without Representation: Labor Issues in Motion and Performance Capture, 11 Animation: An Interdisciplinary J. 40, 46 (2016).

[102] Id.

[103] Dean Goodman & Edwin Chan, Natalie Portman Leaps to Oscar for ‘Black Swan’, Reuters.com (Feb. 27, 2011, 11:20PM), https://www.reuters.com/article/us-oscars-bestactress-idUSTRE71R19920110228.

[104] Mihailova supra note 101, at 46.

[105] See, e.g., Aalmuhammed v. Lee, 202 F.3d 1227, 1232 (9th Cir. 2000)

[106] Id.

[107] Brief of Amici Curiae Professors Shyamkrishna Balganesh et al. Supporting Neither Party, Garcia v. Google, Inc., 786 F.3d 733 (9th Cir. 2015) (No. 12-57302), at 17.

[108] Id. at 13.

[109] Id.

[110] Id. at 13-14.

[111] See generally Homer Croy, How Motion Pictures Are Made, (Goemaere Press 2009) (1918) and J.S. Rudsenske, Music Business Made Simple: A Guide to Becoming a Recording Artist, (J.P. Denk ed. 2004).

[112] See generally Jack Viertel, The Secret Life of the American Musical: How Broadway Shows Are Built (Sarah Crichton Books 2017) (2016).

[113] See generally James R. Martin, Create Documentary Films, Videos, and Multimedia: A Comprehensive Guide to Using Documentary Storytelling Techniques for Film, Video, the Internet and Digital Media Nonfiction Projects (Real Deal Press 2010); Croy, supra note 111.

[114] See Gail Resnik & Scott Trost, All You Need to Know About the Movie and TV Business 36-37 (Simon & Schuster 1996).

[115] Id.

[116] Id.

[117] Balganesh et al., supra note 107, at 14.

[118] See generally Cynthia Littleton & Brian Steinberg, Disney to Buy 21st Century Fox Assets for $52.4 Billion in Historic Hollywood Merger, Variety (Dec. 14, 2017 4:04 AM), http://variety.com/2017/biz/news/disney-fox-merger-deal-52-4-billion-merger-1202631242/.

[119] Id.

[120] Garcia v. Google, Inc., 786 F.3d 733, 736 (9th Cir. 2015).

[121] Id. at 742 (quoting Aalmuhammed v. Lee, 202 F.3d 1227, 1233 (9th Cir. 2000)).

[122] Id.

[123] Id.

[124] Id. at 743.

[125] See generally What is the Writers Guild of America West?, WGA, http://www.wga.org/the-guild/about-us/faq#general1 (last visited Nov. 11, 2018).

[126] See Denise Mann, The Labor Behind the Lost ARG: WGA’s Tentative Foothold in the Digital Age, in Wired TV: Laboring Over an Interactive Future 118, 124 (Denise Mann ed., 2014).

[127] Id. at 118-19.

[128] A mobisode is an episode of television intended for viewing on a mobile device. See generally Scott Ruston, Televisual Narratives in the Palm of Your Hand, Understanding Mobisodes, Producing Television, Spring 2008, https://cinema.usc.edu/assets/054/10994.pdf.

[129] Mann, supra note 126, at 118.

[130] Id. at 135.

[131] Id.

[132] Id. at 119-20

[133] Id. at 122.

[134] Id. at 136.

[135] Id.

[136] Id.

[137] Id.

[138] See Derek Johnson, Authorship Up for Grabs: Decentralized Labor, Licensing, and the Management of Collaborative Creativity, in Wired TV: Laboring Over an Interactive Future 32, 40–41 (Denise Mann ed., 2014).

[139] Id. at 34.

[140] See Writers Guild of America, Screen Credits Manual 16 (2010), http://www.wga.org/contracts/credits/manuals/screen-credits-manual#4 (defining the term “ writer”).

[141] Id.

[142] Catherine L. Fisk, The Role of Private Intellectual Property Rights in Markets for Labor and Ideas: Screen Credit and the Writers Guild of America, 1938-2000, 32 Berkeley J. Emp. & Lab. L. 215, 250 (2011).

[143] Id. at 248.

[144] Id. at 264.

[145] Balganesh et al., supra note 107, at 15-16.

[146] Id. at 16.

[147] Id.

[148] Christopher Jon Sprigman, Remarks for Advanced Copyright Class at New York University School of Law (Spring 2018).

[149] Id.

[150] 17 U.S.C. § 101 (2012).

[151] Cmty. for Creative Non-Violence v. Reid, 490 U.S. 730, 751-52 (1989).

[152] Id. at 752.

[153] Garcia v. Google, Inc., 786 F.3d 733, 742 (9th Cir. 2015).

[154] Anemona Hartocollis, Sued by Harry Potter’s Creator, Lexicographer Breaks Down on the Stand, N.Y. Times, Apr. 16, 2008, at B1.

Think Big! The Need for Patent Rights in the Era of Big Data and Machine Learning

Think Big! The Need for Patent Rights in the Era of Big Data and Machine Learning
By: Hyunjong Ryan Jin*   Download a PDF version of this article here  

Introduction

With AlphaGo’s triumph over the 9-dan Go professional Lee Sedol in March 2016, Google’s DeepMind team conquered the last remaining milestone in board game artificial intelligence.[1] Just nineteen years after IBM Deep Blue’s victory over the Russian chess grandmaster Garry Kasparov,[2] Google’s success exceeded expert predictions by decades.[3] AlphaGo demonstrated how machine learning algorithms could enable processing of vast amounts of data. Played out on a 19 by 19 grid, the number of possible configurations on a Go board is astronomical.[4] With near-infinite number of potential moves, conventional brute-force comparison of all possible outcomes is not feasible.[5] To compete with professional level human Go players, the gaming artificial intelligence requires a more sophisticated approach than the algorithms employed for chess — machine learning. The underlying science and implementation of machine learning was described in a Nature article two months prior to AlphaGo’s match with Lee. In the article, the Google team described how a method called “deep neural networks” decides between the insurmountable number of possible moves in Go.[6] The AlphaGo model was built by reinforcement learning from a database consisting of over thirty million moves of world-class Go players.[7] This allowed the algorithm to optimize the search space of potential moves, therefore reducing the required calculations to determine the next move.[8] In other words, the algorithm mimics human intuition based on the “experience” it gained from the database “fed” into the algorithm, which drastically increases computational efficiency by eliminating moves not worth subsequent consideration. This allows the algorithm to devote computational resources towards the outcomes of “worthwhile” moves. The advent of such powerful analytical tools, capable of mimicking human intuition alongside massive computation power, opens endless possibilities—early stage cancer detection[9], accurate weather forecasting,[10] prediction of corporate bankruptcies,[11] natural event detection,[12] and even prediction of elections.[13] For information technology (“IT”) corporations, investment in such technology is no longer an option, but a necessity. The question that this Note addresses is whether the current state of intellectual property law is adequate to harness the societal benefits that we hope to enjoy through the advances in machine learning. In particular, are patents necessary in the age of big data? And if they are, how should we apply patent protection in the field of big data and machine learning? Part I of this Note examines the need for intellectual property rights in machine learning and identifies the methods by which such protection may be achieved. The differences between trade secret, copyright and patent protection in software are discussed, followed by the scope of protection offered by each means. This background provides the basis to discuss the effectiveness of each method in the context of machine learning and big data innovations. Part II discusses the basics of the underlying engineering principle of machine learning and demonstrates how the different types of intellectual property protection may apply. Innovators may protect their contributions in machine learning by defending three areas—(1) the vast amount of data required to train the machine learning algorithm, (2) innovations in the algorithms itself including advanced mathematical models and faster computational methods, and (3) the resulting machine learning model and the output data sets. Likewise, there are three distinct methods of protecting these intellectual properties: patents, copyright, and secrecy.[14] This Note discusses the effectiveness of each method of intellectual property protection with three principles of machine learning innovation in mind: facilitating data sharing, avoiding barriers to entry from data network effects, and providing incentives to address the key technological challenges of machine learning. This Notes proposes that patents on computational methods adequately balance the concern of patent monopoly and promoting innovation, hence should be the primary means of intellectual property protection in machine learning. Part III then visits the legal doctrine of patentable subject matter starting with the United States Supreme Court’s Alice decision. While Alice imposed a high bar for software patents, the post-Alice Federal Circuit decisions such as Enfish, Bascom, and McRO suggest that certain types of software inventions are still patentable. Specifically, this section will discuss the modern framework pertinent to subject matter analysis: (1) inventions that are directed to improvements of computer functionality rather than an abstract idea, (2) inventions that contain an inventive concept, and (3) inventions that do not improperly preempt other solutions. The Note will apply this framework to innovations in machine learning. The Note proposes that patents for computational methods balance the need for intellectual property protection while permitting data sharing, paving the pathway for promoting innovation in machine learning. The Note further argues that machine learning algorithms are within patentable subject matter under 35 U.S.C. §101.

I. Need for Intellectual Property Rights in Machine Learning

He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening me.” – Thomas Jefferson “I’m going to destroy Android, because it’s a stolen product. I’m willing to go thermonuclear war on this. They are scared to death, because they know they are guilty.” – Steve Jobs The two quotes above demonstrate the conflicting views on protecting intangible ideas with intellectual property law. Thomas Jefferson implied that the free circulation of inventive ideas and thoughts would not dampen the progress of innovation nor disadvantage innovators. On the other hand, Steve Jobs exhibited fury over the similarity between the iOS and the Android OS. Why? Was it because his company was worse off due to the similarity between the two products? Would Apple have refrained from inventing the iPhone had it known others would enter the smartphone market? This section discusses the motives behind the grant of intellectual property rights and whether such protection should be extended to machine learning innovations. Basics of patent law, copyright law, and trade secret are introduced to provide the analytical tools for subsequent discussion on which type of intellectual property protection best promotes the socially-beneficial effects of machine learning.

A. Do We Need Intellectual Property Rights for Machine Learning?

The primary objectives of intellectual property rights are to encourage innovation and to provide the public with the benefits of those innovations.[15] In the context of machine learning, it is not clear whether we need any additional incentives to promote participation in this field. Machine learning is already a “hot field,” with countless actors in industry and academia in active pursuit to keep pace.[16] Hence investment incentivizing may not be a valid justification for granting intellectual property rights in machine learning. Rather, such protection is crucial to promote competition and enhance public benefits. The quality of inferences that may be drawn from a given data set increases exponentially as the aggregation diversifies, which is why cross-industry data aggregation will greatly enhance the societal impact of machine learning.[17] Companies will need to identify new data access points outside of their own fields to gain access to other data sets to further diversity their data. Yet the incentive structures of behemoth corporations may not be well-suited to identify and grow niche markets.[18] It would be up to the smaller, specialized entities to find the gaps that the larger corporations overlooked and provide specialized services addressing the needs of that market. Protective measures that assist newcomers to compete against resource-rich corporations may provide the essential tools for startups to enter such markets. Sufficient intellectual property protection may serve as leverage that startups may use to gain access to data sets in the hands of the Googles and Apples of the world, thus broadening the range of social benefits from machine learning.

B. The Basics of Patent Law

To promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries” – United States Constitution, Article I, § 8 The United States Constitution explicitly authorizes Congress to promote useful arts by granting inventors the exclusive rights of their discoveries. Such constitutional rights stems from two distinct bases — (1) a quid pro quo where the government issues a grant of monopoly in exchange for disclosure to society, and (2) property rights of the inventor. The purpose for such rights is explicitly stated in the Constitution—to promote new inventions. The goal is to prevent second arrivers who have not invested in the creation of the initial invention from producing competing products and services at a lower price, undercutting the innovator whose costs are higher for having invested to create the invention. As an incentive for innovators willing to invest in new, useful arts, the patent system provides the innovator rights to exclude others from practicing the invention. Another purpose of such rights is the concept of “ “mining rights.” Akin to the grant of mining rights to the owner in efforts to suppress aggressive mining, the inventor should have the right to define and develop a given field by excluding other people from the frontiers of that knowledge. Considering the importance of industry standards in modern electronics, such a purpose acknowledges the importance of early stage decisions that may define the trajectory of new technological advances.

C. The Thin Protection on Software Under Copyright Law

The Copyright Act defines a “computer program” as “a set of statements or instructions to be used directly or indirectly in a computer to bring about a certain result.”[19] Though it may be counterintuitive to grant copyright protection for “useful arts” covered by patents, Congress has explicitly mandated copyright protection for software.[20] However, as will be discussed below, copyright protection of software has been significantly limited due to case law. Copyright protects against literal infringement of the text of the program. Source code, code lines that the programmers “author” via computer languages such as C++ and Python, is protected under copyright as literary work.[21] In Apple v. Franklin Corp., the Third Circuit Court of Appeals held that object code, which is the product of compiling the source code, is also considered a literary work.[22] Given that compiled code is a “translation” of the source code, this ruling seems to be an obvious extension of copyright protection. Removing the copyright distinction between source code and object code better reflects the nature of computer languages such as Perl, where the source code is not translated into object code but rather is directly fed into the computer for execution. However, the scope of protection on either type of code is very narrow. The copyright system protects the author against literal copying of code lines. This leaves open the opportunity for competitors to avoid infringement by implementing the same algorithm using different text. Fortunately, in addition to protection against literal copying of code, copyright law may provide some protection of the structure and logical flow of a program. Equivalent to protecting the “plot” of a novel, the Second Circuit Court of Appeals ruled that certain elements of programming structure are considered an expression (copyrightable) rather than idea (not copyrightable), extending copyright protection to non-literal copying.[23] The Computer Associates International v. Altai court applied a three-step test to determine whether a computer program infringes other programs—(1) map levels of abstraction of the program; (2) filter out protectable expression from non-protectable ideas; and (3) compare which parts of the protected expression are also in the infringing program.[24] The merger doctrine is applied to step two of the Altai test to limit what may be protected under copyright law. Under the merger doctrine, code implemented for efficiency reasons is considered as merged with the underlying idea, hence not copyrightable.[25] Since most algorithms are developed and implemented for efficiency concerns, the Altai framework may prevent significant aspects of software algorithms from receiving copyright protection. This means that for algorithms related to computational efficiency, patents may provide significantly more meaningful protection than copyright. The Federal Circuit, in the 2016 case McRO Inc. v. Bandai Namco Games America Inc., ruled that patent claims with “focus on a specific means or method that improves the relevant technology” may still be patentable.[26] Although preemption concerns may impede patentability, exemption of patent right by preemption is narrow compared to that of copyright by the merger doctrine. Scène à faire doctrine establishes yet another limitation on copyright for computer programs. Aspects of the programs that have been dictated by external concerns such as memory limits, industry standards and other requirements are deemed as non-protectable elements.[27] For mobile application software, it is difficult to imagine programs that are not restricted by form factors such as mobile AP computation power, battery concerns, screen size, and RAM limitations. As for machine learning software, the algorithms determine the “worthiness” of computation paths based on conserving computational resources. The external factors that define the very nature and purpose of such machine learning algorithms may exempt them from copyright protection.

D. Comparing Trade Secret and Non-disclosures with Patents

The crucial distinction between trade secret and patent law is secrecy. While patent applicants are required to disclose novel ideas to the public in exchange for a government granted monopoly, trade secret requires owners to keep the information secret. Though trade secret protection prevents outsiders from acquiring the information by improper means, it does not protect the trade secret against independent development or even reverse engineering of the protected information. In trade secret doctrine, the existence of prior disclosed art is only relevant for discerning whether the know-how is generally known, a different and simpler analysis than the issue of novelty in patent law.[28] The United States Supreme Court has specified in Kewanee Oil that all matters may be protected under trade secret law, regardless whether it may or may not be patented.[29] The Kewanee Oil court predicted that inventors would not resort to trade secret when offered a presumptively stronger protection by patent law:
“The possibility that an inventor who believes his invention meets the standards of patentability will sit back, rely on trade secret law, and after one year of use forfeit any right to patent protection, 35 U.S.C. § 102(b), is remote indeed.”[30]
Trade secret is an adequate form of protection for innovators that are concerned with the limits of what may be patentable. The secrecy requirement of trade secret inherently provides protection that may potentially outlive any patent rights, provided a third party does not independently acquire the secret. This coincides with an interesting aspect of machine learning and big data—the need for massive amounts of data. Developers need data to “train” the algorithm, and increase the accuracy of the machine learning models. Companies that have already acquired massive amounts of data may opt to keep their data secret, treating the aggregated data as a trade secret. In addition to the amount of amassed data, companies have all the more reason to keep their data secret if they have access to meaningful, normalized data. Even if a company amasses an enormous amount of data, the data sets may not be compatible with each other. Data gathered from one source may have different reference points or methodologies that are not immediately compatible with data from another source. This raises the concern of “cleaning” massive amounts of data.[31] Such concerns of data compatibility mean that parties with access to a single, homogenous source of high quality data enjoy a significant advantage over parties that need to pull data from multiple sources. However, data secrecy may not be a suitable strategy for companies that are aiming for cross-industry data aggregation. Institutions such as Global Alliance for Genomics and Health are promoting data sharing between research participants. The Chinese e-commerce giant Alibaba announced a data sharing alliance with companies such as Louis Vuitton and Samsung to fight off counterfeit goods.[32] To facilitate the development of technology and to mitigate risks, various companies and research institutions across diverse fields are engaging in joint development efforts and alliances. Seeking protection under trade secret runs against this trend of engaging in effective cross-industry collaboration. Yet there are countervailing arguments that trade secret promotes disclosure by providing legal remedies that can replace the protection of secrets.[33] Parties can sidestep the limitations of trade secrets by sharing proprietary information under the protection of contract law. While data sharing practices may void trade secret protection, the nature of continued accumulation of data and carefully drafted contractual provisions may provide sufficient protection for the data owners.

II. Placing Machine Learning within Intellectual Property Law

Learning is any process by which a system improves performance from experience.” – Herbert Simon, Nobel Prize in Economics 1978. The concept of machine learning relates to computer programs that have the capability to improve performance based on experience, with limited intervention of the programmer.[34] Machine learning models have the capability to automatically adapt and customize for individual users, discover new patterns and correlations from large databases, and automate tasks that require some intelligence by mimicking human intuition.[35] This section dissects the mechanics of machine learning to identify the aspects of machine learning innovations that are at issue as intellectual property.

A. Machine Learning Basics

Machine learning methods are divided into two different approaches—supervised machine learning and unsupervised machine learning. For supervised machine learning, models are typically established by applying “labeled” sets of data to a learning algorithm. Labeled data refers to data sets that have both relevant features and the target results that the programmer is interested in. For example, we may be interested in developing a machine learning model that classifies images with dogs in them. The data sets for supervised machine learning would indicate whether a given images has dogs or not. The learning process begins with the algorithm fitting trends found in the training data set into different types of models. The algorithm compares the prediction errors of the models by inputting the validation set data into each model, measuring their accuracy. This allows the algorithm to decide which of the various models is best suited as the resulting machine learning model. Finally, the machine learning model is then evaluated by assessing the accuracy of the predictive power of the model. The developed model is then applied to data without a correct answer to test the validity of the model. In unsupervised machine learning, the data sets are “unlabeled” data, which may not contain the result that the programmer is interested in. Returning to our dog image classification example, data sets for unsupervised machine learning will have pictures of various animals that are not labeled—the computer does not know which pictures are associated with dogs. The unsupervised machine learning algorithm develops a model that extracts common elements from the picture, teaching itself the set of features that makes the subject of the picture a dog. In essence, unsupervised machine learning uses data sets that do not have specific labels fed into the algorithm for the purpose of identifying common trends embedded in that data set. The objective of developing such machine learning models varies. Sometimes the goal is to develop a prediction model that can forecast a variable from a data set. Classification, which assigns records to a predefined group, is also a key application of the algorithm. Clustering refers to splitting records into distinct groups based on the similarity within such group. Association learning identifies the relationship between features.         Figure 1. Overview of Machine Learning Model Development   Figure 1 illustrates the overall process of machine learning model development. The learning process of machine learning algorithms begins with aggregation of data. The data originates from an array of diverse sources ranging from user input, sensor measurement, or monitoring of user behavior.[36] The data sets are then preprocessed. The quality of data presents a challenge in improving machine learning models—any data that has been manually entered contains the possibility of error and bias.[37] Even if the data is collected through automatic means, such as health monitoring systems or direct tracking of user actions, the data sets require preprocessing to account for systematic errors associated with the recording device or method.[38] This includes data skews due to difference between individual sensors, errors in the recording or transmission of data, and incorrect metadata about the sensor.[39] Simply put, the data sets may have differing reference points, embedded biases, or differing formats. The “cleaning” process accommodates for the data skews. The objective of machine learning models is to identify and quantify “features” from a given data set. The term “feature” refers to individually measurable property of an observed variable.[40] From the outset, there may be an extensive list of features that are present in a set of data. It would be computationally expensive to define and quantify each feature, and then to identify the inter-feature relationships, from massive amounts of data. Due to the high demand for the computational power required for processing massive amounts of data, dedication of computational resources to features that are outside the scope of the designer’s interest would be a waste of such limited computational capacity.[41] The machine learning algorithm reduces waste of computational resources by applying dimensionality reduction to the pre-processed data sets.[42] The algorithm can identify an optimal subset of features by reducing the dimension and the noise of the data sets.[43] Dimensionality reduction allows the machine learning model to achieve higher level of predictive accuracy, increased speed of learning, and improves the simplicity and comprehensibility of the results.[44] However, the reduction process has limitations—reducing dimensionality inevitably imposes a limit on the amount of insights and information that may be extracted from the data sets. If the machine learning algorithm discerns a certain feature, the model would not be able to draw inferences related to said feature. Following dimensionality reduction, the machine learning algorithm attempts to fit the data sets into preset models. Typically, three different types of data are fed into the machine learning model—training set, validation set, and test set.[45] The machine learning algorithm “trains” the model by fitting the training set data into various models to evaluate the accuracy of each selection. Then the validation set is used to estimate error rates of each model when applied to data outside the training set that was used to develop each model. Through this process, the machine learning algorithm selects the model that best describes the characteristics and trends of the target features from the test and validation sets.[46] The test set is then used to calculate the generalized prediction error, which is reported to the end user for proper assessment of the predictive power of the model.[47] Simply put, the training test and validation set is used to develop and select a model that reflects the trends of the given data set, and the test set is used to generate a report on the accuracy of the selected model. The crucial elements in developing a machine learning model are (1) training data, (2) inventions related to the machine learning algorithm such as the method of preprocessing the training data, the method of dimensional reduction, feature extraction, and the method of model learning/testing, and (3) the machine learning model and output data.[48] An ancillary element associated with the three elements above is the human talent that is required to implement such innovation.[49] Innovators in the field of machine learning may protect their investments by protecting one or more of the elements listed above. The difference between training data and output data, as well as the difference between the machine learning algorithm and the machine learning model, are best illustrated with an example. Let us assume a credit card company wants to use machine learning to determine whether the company should grant a premium credit card to a customer. Let us further assume that the company would prefer to grant this card to customers that would be profitable to the company while filtering out applicants that are likely to file for bankruptcy. Data sets about prior applicant information would correspond to training data. The company would apply a mathematical method of extracting insight about the correlation between features and the criteria that the company wants to evaluate (e.g., profitable for the firm or likely to file bankruptcy). The mathematical methods are referred as machine learning algorithms. The resulting mechanism, such as a scoring system, that determines the eligibility of card membership is the machine learning model. The credit card applicant’s personal data would be the input data for the machine learning model, and the output data would include information such as expected profitability of this applicant and likelihood of bankruptcy for this applicant.

B. Industry Trends in Machine Learning

Discussing incentive structures and trends behind the machine learning industry is essential in identifying adequate methods of intellectual property rights. The current trends in the world of machine learning will predict what intellectual property regime is most useful to companies to protect their work. The United States has chronically struggled to maintain adequate supply of talent in the high-tech industry, a deficit of talent that continues in the field of machine learning.[50] From a report by the McKinsey Global Institute, the United States’ demand for talent in deep learning “could be 50 to 60 percent greater than its projected supply by 2018.”[51] Coupled with the dearth of machine learning specialists, the short employment tenure of software companies further complicates the search for talent. Software engineers from companies such as Amazon and Google have reported an average employment tenure of one year.[52] While some parts of the high attrition rate may be attributed to cultural aspects of the so-called “Gen Y” employees, the “hot” demand for programming talent has significant impact on the short employee tenure.[53] Job mobility within the software industry is likely to increase as the “talent war” for data scientist intensifies. Employee mobility and California’s prohibition against “covenants not to compete” have been accredited as a key factor behind the success of Silicon Valley.[54] Another trend in the field is the rapid advances in machine learning methods. Due to the fast-paced development of the field, data scientists and practitioners have every reason to work with companies that would allow them to work at the cutting edge of machine learning, using the best data sets. This may influence the attrition rates and recruiting practices of the software industry mentioned above.[55] Eagerness of employees to publish scientific articles and contribute to the general machine learning committee may be another factor of concern. To accelerate innovation by repurposing big data for uses different from the original purpose, and to form common standards for machine learning, more industries are joining alliances and collaborations.[56] Cross-industry collaborations may enable endless possibilities. Imagine the inferences that may be drawn by applying machine learning methods to dietary data from home appliances, biometric data, and data on the weather patterns around the user. Putting privacy nightmares aside, machine learning with diverse data sets may unlock applications that were not previously possible. More companies are attempting to capitalize on commercial possibilities that data sharing may unlock.[57]

C. Machine Learning Innovators – Protect the Data or Inventions?

Though it may seem intuitive that patent protection may be the best option, innovations in machine learning may not need patent protection. Trade secret protection on the data sets may be sufficient to protect the interests of practicing entities while avoiding disclosure of their inventions during the patent prosecution process. Furthermore, numerous software patents have been challenged as unpatentable abstract subject matter under 35 U.S.C. §101 since the Alice decision in 2014.[58] Though subsequent decisions provided guidelines for types of software patents that would survive the Alice decision, it is not clear how the judiciary will view future machine learning patents. Such issues raise the question about the patentability of machine learning – should we, and can we, resort to patents to protect machine learning inventions? Following the discussion on the building blocks of machine learning and recent emerging trends in the field, this section discusses the mode and scope of protection that current legal system provides for each element pertinent to innovation in machine learning. The possible options for protecting innovations are (1) non-disclosure agreements and trade secret law, (2) patent law, and (3) copyright. The three options for protection may be applied to the three primary areas of innovation—(1) training data, (2) inventions related to computation, data processing, and machine learning algorithms, and (3) machine learning models and output data. This discussion will provide context about the methods of protection for innovations in machine learning by examining the costs and benefits of the various approaches.
1. Protecting the Training Data—Secrecy Works Best
Access to massive amounts of training data is a prime asset for companies in the realm of machine learning. The big data phenomenon, which triggered the surge of interest in machine learning, is predicated on the need for practices to analyze large data resources and the potential advantages from such analysis.[59] Lack of access to a critical mass of training data prevents innovators from making effective use of machine learning algorithms. Previous studies suggest that companies resent sharing data with each other.[60] Michael Mattioli discusses the hurdles against sharing data and considerations involved with reuse of data in his article Disclosing Big Data.[61] Indeed, there may be practical issues that prevent recipients of data from engaging in data sharing. Technical challenges in comparing data from different sources, or inherent biases embedded in data sets may be reasons that complicate receiving outside data.[62] Mattioli also questions the adequacy of the current patent and copyright system to promote data sharing and data reuse—information providers may prefer not to disclose any parts of their data due to the rather thin legal protection for databases.[63] Perhaps this is why secrecy seems to be the primary method of protecting data.[64] The difficulty of reverse engineering to uncover the underlying data sets promotes the reliance on non-disclosure.[65] Compared to the affirmative steps required to maintain trade secret protection if the data is disclosed, complete non-disclosure may be a cost effective method of protecting data.[66] Companies that must share data with external entities may exhibit higher reliance on contract law rather than trade secret law. In absence of contract provisions, it would be a challenge to prove that the trade secret has been acquired by misappropriation of the recipient party. The “talent war” for data scientists may also motivate companies to keep the training data sets secret. With a shortage of talent to implement machine learning practices and rapid developments in the field, retaining talent is another motivation for protecting against unrestricted access to massive amounts of data. Companies may prefer exclusivity to the data sets that programmers can work with — top talents in machine learning are lured to companies with promises of exclusive opportunities to work with massive amounts of data.[67] The rapid pace of development in this field encourages practitioners to seek opportunities that provide the best resources to develop their skill sets. This approach is effective since a key limitation against exploring new techniques in this field is the lack of access to high quality big data. Overall, secrecy over training data fits well with corporate recruiting strategies to retain the best talents in machine learning. Non-disclosure and trade secret protection seems to be the best mode of protection. First, despite the additional legal requirements necessary to qualify as trade secrets, trade secret protection fits very well with non-disclosure strategy. On the other hand, patent law is at odds with the principle of non-disclosure. While trade secret law provides companies protection without disclosing information, patent law requires disclosure in exchange for monopolistic rights. Furthermore, neither patent nor copyright provide adequate protection for underlying data. Patent law rewards creative concepts and inventions, not compiled facts themselves. Copyright may protect labeling or distinct ways of compiling information, but does not protect underlying facts. Also, as a practical matter, the difficulty of reverse engineering of machine learning models does not lend well to detecting infringement. Analysis of whether two parties used identical training data would not only be time consuming and costly, but may be fundamentally impossible. If companies were to seek protection of training data, it would be best to opt for secrecy by non-disclosure. This would mean companies would opt out of the cross-industry collaborations that were illustrated above. This may be less of a concern for innovation, as companies may still exchange output data as means of facilitating cross-industry collaboration.
2. Protecting the Inventions—Patent Rights Prevail
Adequate protection over inventive approaches in processing data is becoming increasingly important as various industries begin to adopt a collaborative alliance approach in machine learning. Cross-industry collaboration requires implementation of methods such as preprocessing diverse data sets for compatibility. As the sheer amount of data increases, more processing power is required. The machine learning algorithm needs to maintain a high degree of dimensionality to accurately identify the correlations between a high number of relevant features. The need for more innovative ideas to address such technological roadblocks will only intensify as we seek more complex applications for machine learning. The three primary areas where novel ideas would facilitate innovations in machine learning are pre-training data processing, dimensional reduction, and the machine learning algorithm. Access to massive amounts of data alone is not sufficient to sustain innovation in machine learning. The raw data sets may not be compatible with each other, requiring additional “cleaning” of data prior to machine learning training.[68] The data provided to the machine learning algorithm dictates the result of the machine learning model, hence innovations in methods to merge data with diverse formats is essential to enhancing the accuracy of the models. As cross-industry data analysis becomes more prominent, methods of merging data will have more significant impact on advancing the field of machine learning than mere collection of large data sets. Cross-industry data sharing would be useless unless such data sets are merged in a comparable manner.[69] Companies can opt to protect their inventive methods by resorting to trade secret law. The difficulty of reverse engineering machine learning inventions, coupled with the difficulty of patenting software methods provides incentives for innovators to keep such inventions secret from the public. However, two factors would render reliance on non-disclosure and trade secret ineffective—frequent turnover of software engineers and rapid speed of development in the field. Rapid dissemination of information from employment mobility may endanger intellectual property protection based on secrecy. Furthermore, while the law will not protect former employees that reveal trade secrets to their new employers, the aforementioned fluid job market coupled with general dissemination of information make it difficult to distinguish between trade secrets from former employment and general knowledge learned through practice. The difficulties of reverse engineering machine learning models work against the trade secret owner as well in identifying trade secret misappropriation—how do you know others are using your secret invention? The desire for software communities to discuss and share recent developments in the field does not align well with the use of secrecy against innovations in machine learning. Secrecy practices disincentivize young data scientists from joining due to the limits against gaining recognition.[70] The rapid development of machine learning technology also presents challenges against reliance on trade secret law. Secret methods may be independently developed by other parties. Neither trade secret law nor non-disclosure agreements protect against independent development of the same underlying invention.[71] Unlike training data, machine learning models, or the output data, there are no practical limitations that impedes competitors from independently inventing new computational methods of machine learning algorithms. With such a fluid employment market, high degree of dissemination of expertise, and rapid pace of development, patent protection may provide the assurance of intellectual property protection for companies developing inventive methods in machine learning. Discussions on overcoming the barriers of patenting software will be presented in later sections.[72]
3. Protecting the Machine Learning Models and Results—Secrecy Again
The two primary products from applying the machine learning algorithms to the training data are the machine learning model and the accumulation of results produced by inputting data into the machine learning model. The “input data” in this context may refer to individual data that is analyzed by the insights gained from the machine learning model. In a recent article, Brenda Simon and Ted Sichelman discuss the concerns of granting patent protection for “data-generating patents,” which refers to inventions that generate valuable information in their operation or use.[73] Exclusivity based on patent protection may be extended further by trade secret protection over the data that has been generated by the patented invention.[74] Simon and Sichelman argue that the extended monopoly over data may potentially overcompensate inventors since the “additional protection was not contemplated by the patent system[.]”[75] Such expansive rights will cause excessive negative impact on downstream innovation and impose exorbitant deadweight losses.[76] The added protection over the resulting data derails the policy rationale behind the quid pro quo exchange between the patent holder and the public by excluding the patented information from public domain beyond the patent expiration date.[77] The concerns addressed in data-generating patents also apply to machine learning models and output data. Corporations may obtain patent protection over the machine learning models. Akin to a preference for secrecy for training data, non-disclosure would be the preferred mode of protection for the output data. The combined effect of the two may lead to data network effects where users have strong incentives to continue the use of a given service.[78] The companies that have exclusive rights over the machine learning model and output data gather more training data, increasing the accuracy of their machine learning products. The reinforcement by monopoly over the means of generating data allows few companies to have disproportionately strong dominance over their competitors.[79] Market dominance by data-generating patents becomes particularly disturbing when the patent on a machine learning model preempts other methods in the application of interest. Trade secret law does not provide protection against independent development. However, if there is only one specific method to obtain the best output data, no other party would be able to create the output data independently. The exclusive rights over the only methods of producing data provides means for the patent holder to monopolize both the patent and the output data.[80] From a policy perspective, the excessive protection does seem troubling. Yet such draconian combinations are less feasible after the recent rulings on patentable subject matter of software, which will be discussed below.[81] Mathematical equations or concepts are likely directed to an “abstract concept,” thus will be deemed directed to a patent ineligible subject matter.[82] Furthermore, though recent cases in the Federal Circuit have found precedents where software patents passed the patentable subject matter requirement, those cases expressed limitations against granting patents that would improperly preempt all solutions to a particular problem.[83] The rapid pace of innovation in the field of machine learning compared to the rather lengthy period required to obtain patents may also dissuade companies from seeking patents. Overall, companies have compelling incentives to rely on non-disclosure and trade secrets to protect their machine learning models instead of seeking patents. The secrecy concerns regarding training data applies to machine learning models and the output data as well. Non-disclosure would be the preferred route of obtaining protection over the two categories. However, use of non-disclosure or trade secrets to protect machine learning models and output data presents challenges that are not present in the protection of training data. The use of secrecy to protect machine learning models or output data conflicts with recruiting strategies to hire and retain top talent in the machine learning field. The non-disclosure agreements limit the employee’s opportunity to gain recognition in the greater machine learning community. In a rapidly developing field where companies are having difficulty hiring talent, potential employees would not look fondly on corporate practices that limit avenues of building a reputation within the industry.[84] Companies have additional incentives to employ a rather lenient secrecy policy for machine learning models and the output data. They have incentives to try to build coalitions with other companies to monetize on the results. Such cross-industry collaboration may be additional source of income for those companies. The data and know-how that Twitter has about fraudulent accounts within their network may aid financial institutions such as Chase with novel means of preventing wire fraud. The reuse of insights harvested from the large amount of raw training data can become a core product the companies would want to commercialize. Data reuse may have an incredible impact even for applications ancillary to the primary business of the company. Interesting aspects of disclosing machine learning models and output data are the difficulty of reverse engineering and consistent updates. If the company already has sufficient protection over the training data and/or the computational innovations, competitors will not be able to reverse engineer the machine learning model from the output data. Even with the machine learning model, competitors will not be able to provide updates or refinements to the model without the computational techniques and the sufficient data for training the machine learning algorithm. In certain cases, the result data becomes training data for different applications, which raises concerns of competitors using the result data to compete with the innovator. Yet the output data would contain less features and insights compared to the raw training data that the innovator possesses, and therefore would inherently be at a disadvantage when competing in fields that the innovator has already amassed sufficient training data. Grant of patents on machine learning models may incentivize companies to build an excessive data network while preempting competitors from entering competition. This may not be feasible in the future, as technological preemption is becoming a factor of consideration in the patentable subject matter doctrine. Companies may use secrecy as an alternative, yet may have less incentives to keep secrecy compared to the protection of training data.

D. Need of Patent Rights for Machine Learning Inventions in the Era of Big Data

The current system, on its surface, does not provide adequate encouragement for data sharing. If anything, companies have strong incentives to avoid disclosure of their training data, machine learning model, and output data. Despite these concerns, data reuse may enable social impacts and advances that would not be otherwise possible. Previous studies have pointed out that one of the major barriers preventing advances in machine learning is the lack of data sharing between institutions and industries.[85] Data scientists have demonstrated that they were able to predict flu trends with data extracted from Twitter.[86] Foursquare’s location database provides Uber with the requisite data to pinpoint the location of users based on venue names instead of addresses.[87] Information about fraudulent Twitter accounts may enable early detection of financial frauds.[88] The possibilities that cross-industry data sharing may bring are endless. To encourage free sharing of data, companies should have a reliable method of protecting their investments in machine learning. At the same time, protection based on non-disclosure of data would defeat of purpose of promoting data sharing. Hence protection over computation methods involved with machine learning maintains the delicate balance between promoting data sharing and protecting innovation. Protection over inventions in the machine learning algorithm provides one additional merit other than allowing data sharing and avoiding the sort of excessive protection that leads to a competitor-free road and data network effects. It incentivizes innovators to focus on the core technological blocks to the advancement of technology, and encourages disclosure of such know-how to the machine learning community. Then what are the key obstacles in obtaining patents in machine learning inventions? While there are arguments that the definiteness requirement of patent law is the primary hurdle against patent protection of machine learning models due to reliance on subjective judgment, there is no evidence that the underlying inventions driving big data faces the same challenge.[89] Definiteness may be overcome by providing reasonable certainty for those skilled in the art of defining what the scope of the invention is at the time of filing.[90] There is no inherent reason why specific solutions for data cleaning, enhancement of computation efficiency, and similar inventions would be deemed indefinite by nature. Since the United States Supreme Court invalidated a patent on computer implemented financial transaction methods in the 2014 Alice decision, the validity of numerous software and business method patents were challenged under 35 U.S.C. §101.[91] As of June 8th, 2016, federal district courts invalidated 163 of the 247 patents that were considered under patentable subject matter—striking down 66% of challenged patents.[92] The U.S. Court of Appeals for the Federal Circuit invalidated 38 of the 40 cases it heard.[93] Arguably, the public benefits more from such high rates of post-issuance invalidity. The public still has access to the disclosures from the patents and patent applications. In reliance on granted patents, companies may have already invested in growing related businesses, catering to the need of consumers. At the same time, the patent holder’s monopolistic rights have been shortened as the result of litigation. Effectively, the price that the public pays to inventors in exchange for the benefits of disclosure is reduced. Yet the high degree of invalidity raises several concerns for the software industry. Smaller entities, lacking market influence and capital, have difficulty competing against established corporations without the monopolistic rights granted through the patent system. Investors become hesitant to infuse capital into startups for fear that invalidity decreases the worth of patents. Reliance on trade secret has its own limitations due to the disclosure dilemma—the inventor needs to disclose the secret to lure inventors, but risks losing secrecy in the process. Copyright law does not provide appropriate protection. The restrictions imposed by the merger doctrine and scène à faire doctrine constrain copyright protection of software. Though copyright provides an alternative method of protecting literal copying of code, it does little to protect the underlying software algorithms and innovation. Ultimately, the increase of alliances and collaboration provides incentives for parties to obtain patent rights. Reliance on trade secret or copyright are not suitable methods of protecting their intellectual property. Furthermore, market power or network effects alone cannot sufficiently mitigate the risks involved with operating a business. Patents become even more important for startups since patents provide investors with assurance that in the worst case, the patents may still serve as potential collateral.        

III. Patentability of Machine Learning Innovations in the Era of Big Data

  Patentable subject matter continues to be a barrier for patenting innovations in software. Additional doctrines such as enablement, written description, and obviousness are also serious obstacles against obtaining patents, yet such requirements are specific to each claimed invention and the draftsmanship of claims. Subject matter is considered a broader, categorical exclusion of patent rights. This section explores the current landscape of the patentable subject matter doctrine in the software context.

A. Alice: The Legal Framework of Patentable Subject Matter in Software

The complexity involved with software, coupled with the relatively broad scope of software patents, has presented challenges in identifying the boundaries of the claims.[94] Many members of the software community detest imposing restrictions on open source material and attest that many key innovations in algorithms are rather abstract.[95] Such hostility against patenting software has raised the question of whether patent rights should be the proper method of protecting innovations in software. Alice was a case that embodied such opposition to the grant of software patents. The case involved patents on computerized methods for financial trading systems that reduce “settlement risk” when only one party to financial exchange agreement satisfies its obligation.[96] The method proposed the use of a computer system as a third-party intermediary to facilitate the financial obligations between parties.[97] The United States Supreme Court ruled that the two-step test established from Mayo governed all patentable subject matter questions.[98] In particular, for the abstract idea context, the Supreme Court established the following two-step framework for patentable subject matter of software inventions:  
1. Step one: “[D]etermine whether the claims at issue are directed to a patent-ineligible concept. If so, the Court then asks whether the claim’s [additional] elements, considered both individually and ‘as an ordered combination,’ ‘transform the nature of the claim’ into a patent-eligible application.”[99]
 
2. Step two: “[E]xamine the elements of the claim to determine whether it contains an ‘inventive concept’ sufficient to ‘transform’ the claimed abstract idea into a patent-eligible application. A claim that recites an abstract idea must include ‘additional features’ to ensure that the [claim] is more than a drafting effort designed to monopolized the [abstract idea]” which requires “more than simply stat[ing] the [abstract idea] while adding the words ‘apply it.’”[100]
  The Alice Court found that the patent on financial transaction was “directed to a patent-ineligible concept: the abstract idea of intermediated settlement,” and therefore failed step one.[101] Furthermore, the Court ruled that the claims did “no more than simply instruct the practitioner to implement the abstract idea of intermediated settlement on a generic computer” and did not provide an inventive concept that was sufficient to pass step two.[102]

B. The post-Alice cases from the Federal Circuit

The Alice framework was considered as a huge setback for the application of patentable subject matter doctrine to software. It was a broad, categorical exclusion of certain inventions that were deemed “directed to” an abstract idea, natural phenomenon, or law of nature. The biggest misfortune was the lack of guidance in the Alice decision on the threshold for such categorical exclusion—we were left without any suggestions on the type of software patents that would be deemed as patentable subject matter. The recent line of cases in the Federal Circuit provides the software industry with the much-needed clarification on the standards that govern patentability of software inventions. Enfish v. Microsoft, decided on March 2016, involved a “model of data for a computer database explaining how the various elements of information are related to one another” for computer databases.[103] In June 2016, the Federal Circuit decided another case on the abstract idea category for patentable subject matter. Bascom Global v. AT&T Mobility is on a patent disclosing an internet content filtering system located on a remote internet service provider (ISP) server.[104] Shortly after Bascom, the Federal Circuit decided McRO v. Bandai Namco Games in September 2016.[105] The case ruled that an automated 3D animation algorithm that renders graphics in between two target facial expressions is patentable subject matter.[106] The rulings from the Federal Circuit on the aforementioned three cases provide guidelines along the two-step Alice test of patentable subject matter. The software patents in Enfish and McRO were deemed “directed to” a patent eligible subject matter, informing the public of what may pass the first set of the Alice test. Bascom failed the first step.[107] Yet the court ruled that those patents had inventive concepts sufficient to transform a patent ineligible subject matter into a patent eligible application. Combined together, the three cases give more certainty in what may pass the 35 U.S.C. §101 patentable subject matter inquiry. Reiterating the Alice test, whether an invention is a patentable subject matter is determined by a two-step process—(1) is the invention directed to, rather than an application of, an abstract idea, natural phenomenon, or law of nature, and even if so, (2) do the elements of the claim, both individually and combined, contain an inventive concept that transforms this invention into a patent-eligible application? The Federal Circuit fills in the gaps that were left unexplained from the Alice ruling.
1. The Federal Circuit’s Standard for Alice Step One
The Enfish court discussed what constitutes an abstract idea at the first step of the Alice inquiry. Judge Hughes instructs us to look at whether the claims are directed to a specific improvement rather than an abstract idea. In this case, the patent provides the public with a solution to an existing problem by a specific, non-generic improvement to computer functionality. The Enfish court ruled that such invention is patent eligible subject matter.[108] McRO also ruled that the facial graphic rendering for 3D animation was not an abstract concept. Here, the Federal Circuit again emphasized that a patent may pass step one of the Alice test if the claims of the patent “focus on a specific means or method that improves the relevant technology.”[109] The McRO court also noted that preemption concerns may be an important factor for the 35 U.S.C. §101 subject matter inquiry—that improper monopolization of “the basic tools of scientific and technological work” is a reason why such categorical carve outs against granting patents on abstract ideas exist.[110] Bascom provides the standards on what would fail step one of the Alice patentable subject matter inquiry. If the patent covers a conventional, well-known method in the field of interest, then the invention would be considered abstract. This is akin to the inventive concept considerations conducted at the second phase of the 35 U.S.C. §101 subject matter inquiry. The main takeaway from Enfish and McRO is that in the first step of the Alice test, a patent application is not directed to an abstract idea if (1) the invention addresses an existing problem by specific improvements rather than by conventional, well-known methods and (2) the claims do not raise preemption concerns. This encourages practitioners to define the problem as broadly as possible, while defining the scope of improvement in definite terms.
2. The Federal Circuit’s Standard for Alice Step Two, and the Overlap with Step One
The second step of the Alice test is an inquiry of whether the patent application, which is directed to a patent ineligible subject, still contains a patent-worthy inventive concept. Bascom ruled in favor of granting the patent following the second step of the Alice test.[111] While the patent at hand was considered directed to patent ineligible subject matter, the Bascom court found that the content filter system invention still had an inventive concept worthy of a patent.[112] Even if elements of a claim are separately known in prior art, an inventive concept can be found in the non-conventional and non-generic arrangement of known, conventional pieces. This inquiry seems like a lenient standard compared to the 35 U.S.C. §103 obviousness inquiry; hence, it is not clear if this step has an independent utility for invalidating or rejecting a patent. Nonetheless, the court found that merely showing that all elements of a claim were already disclosed in prior art was not sufficient reason to make an invention patent ineligible.[113] While it is possible to infer sufficient reasons of ruling out inventive concepts from the Bascom case, it is still unclear what would warrant an invention to pass the second step of the Alice test. Cases such as DDR Holdings v. Hotels.com have suggested that the second step of Alice is satisfied since it involved a solution to a specific technological problem that “is necessarily rooted in computer technology in order to overcome a problem specifically arising in the realm of computer networks.”[114] This interpretation of inventive concept becomes perplexing when comparing the two steps of Alice—both steps look to whether the proposed solution addresses problems that are specific to a given field of interest. While we would need additional cases to gain insight on whether the two steps have truly distinct functions, at the very least the Federal Circuit provided essential guidelines on what may be deemed as patentable software.

C. Applying Patentable Subject Matter to Machine Learning Inventions

As the Bascom court has taught, the first step in the Alice inquiry is to ask whether an invention (1) provides a solution to an existing problem by (2) a specific, non-generic improvement that (3) does not preempt other methods of solving the existing problem. Applying this test to inventions in machine learning, mathematical improvements and computational improvements would be treated differently. As mentioned before, a key aspect of machine learning is the “noise” associated with the data sets.[115] Another concern is the fitting of a given algorithm to a certain model. Methods that facilitate the computations of the training process may be deemed as a specific improvement. However, machine learning algorithms themselves, including the base models that the algorithm fits the training into would not be pertinent to just a specific improvement. Hence, generic mathematical methods applicable to various problems are directed to an abstract idea. For example, an invention that addresses the issue of normalizing data from different sources would be a computational issue and hence would pass the Alice test given that it did not preempt other solutions to the problem of data normalization. On the other hand, a specific mathematical equation that serves as a starting model for the machine learning algorithm would be mathematical and hence directed to an abstract idea. Even if the mathematical starting model is only good for a specific application, the model is not a specific improvement pertinent to that application. Although the model may not necessarily be a good starting model for other applications, it is nonetheless a generic solution that applies to other applications as well.

Conclusion

While highly restrictive, the guidelines from the Federal Circuit still allow the grant of patent rights for the computational aspects of machine learning algorithms. The guidelines also would prevent highly preemptive mathematical innovations, including data-generating patents such as machine learning models. The narrow range of patentability makes a patent regime appealing for computational methods. The recent emphasis on preemption concerns acts in favor of preventing data network effects based on data-generating patents. While not discussed in this paper, other patentability requirements such as obviousness or definiteness would further constraint the grant of overly broad data-generating patents. Such an approach strikes the appropriate balance between promoting innovation and encouraging data reuse for societal benefits. Compared to other approaches of providing protection over innovations in machine learning, the narrowly tailored approach for patent rights for computational inventions fits best with the policy goal of promoting innovation through data reuse. The industry trends in collaboration and recruiting also matches the proposed focus on patent law protection.
* J.D. Candidate, New York University School of Law, 2018; Ph.D. in Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, 2012. The author would like to thank Professor Katherine Jo Strandburg for her guidance; fellow JIPEL Notes Program participants Julian Pymento, Gia Wakil, Neil Yap, and Vincent Honrubia for their comments and feedback; and Dr. Sung Jin Park for her support throughout the process.
[1] Sang-Hun Choe & John Markoff, Master of Go Board Game Is Walloped by Google Computer Program, N.Y. Times (March 9, 2016), https://www.nytimes.com/2016/03/10/world/asia/google-alphago-lee-se-dol.html (reporting the shocking defeat of Go Master Lee Se-dol to Google DeepMind’s AlphaGo).
[2] Laurence Zuckerman, Chess Triumph Gives IBM a Shot in the Arm, N.Y. Times (May 12, 1997), http://politics.nytimes.com/library/cyber/week/051297ibm.html (detailing IBM’s highly publicized win through Deep Blue’s victory over world chess champion Garry Kasparov).
[3] See Choe & Markoff, supra note 1.
[4] David Silver et al., Mastering the game of Go with deep neural networks and tree search, 529 Nature 484, 484 (2016).
[5] Id.
[6] Id.
[7] Id. at 485.
[8] Id.
[9] See Andre Esteva et al., Dermatologist-level classification of skin cancer with deep neural networks, 542 Nature 115 (2017).
[10] See Sue Ellen Haupt & Branko Kosovic, Big Data and Machine Learning for Applied Weather Forecasts, (2015).
[11] See Wei-Yang Lin et al., Machine Learning in Financial Crisis Prediction: A Survey, 42 IEEE Transactions on Systems, Man, and Cybernetics 421 (2012).
[12] See Farzindar Atefeh & Wael Khreich, A Survey of Techniques for Event Detection in Twitter, 31 Computational Intelligence 132 (February 2015).
[13] See Corey Blumenthal, ECE Illinois Students Accurately Predicted Trump’s Victory, ECE Illinois (Nov. 18, 2016), https://www.ece.illinois.edu/newsroom/article/19754.
[14] For the purpose of this Note, secrecy refers to the use of trade secret and contract based non-disclosure agreements.
[15] Mark A. Lemley, The Surprising Virtues of Treating Trade Secrets As IP Rights, 61 Stan. L. Rev. 311, 332 (2008) (“Patent and copyright law do not exist solely to encourage invention, however. A second purpose — some argue the main one — is to ensure that the public receives the benefit of those inventions.”).
[16] Andrew Ng et al., How Artificial Intelligence Will Change Everything, Wall Street Journal (March 7, 2017), https://www.wsj.com/articles/how-artificial-intelligence-will-change-everything-1488856320.
[17] Limor Peer, Mind the Gap in Data Reuse: Sharing Data Is Necessary But Not Sufficient for Future Reuse, London Sch. Econ. & Poli. Sci. (Mar. 28, 2014) http://blogs.lse.ac.uk/impactofsocialsciences/2014/03/28/mind-the-gap-in-data-reuse (“The idea that the data will be used by unspecified people, in unspecified ways, at unspecified times . . . is thought to have broad benefits”).
[18] See Saeed Ahmadiani & Shekoufeh Nikfar, Challenges of Access to Medicine and The Responsibility of Pharmaceutical Companies: A Legal Perspective, 24 DARU Journal of Pharmaceutical Sciences 13 (2016) (discussing how “pharmaceutical companies find no incentive to invest on research and development of new medicine specified for a limited population . . .”).
[19] 17 U.S.C. §101 (2012).
[20] Id.
[21] 17 U.S.C. §102(a) (Copyright exists “in original works of authorship fixed in any tangible medium of expression . . .”).
[22] Apple Comput., Inc. v. Franklin Comput. Corp., 714 F.2d 1240 (3d Cir. 1983).
[23] Comput. Assocs. Int’l v. Altai, 982 F.2d 693 (2d Cir. 1992).
[24] Id.
[25] See id. at 707-09.
[26] 837 F.3d 1299, 1314 (Fed. Cir. 2016).
[27] Altai, 982 F.2d at 698.
[28] See Dionne v. Se. Foam Converting & Packaging, Inc., 240 Va. 297 (1990).
[29] Kewanee Oil v. Bicron Corp., 416 U.S. 470 (1974).
[30] Id. at 490.
[31] Nikolay Golova & Lars Rönnbäck, Big Data Normalization For Massively Parallel Processing Databases, 54 Computer Standards & Interfaces 86, 87 (2017).
[32] Jon Russell, Alibaba Teams Up with Samsung, Louis Vuitton and Other Brands to Fight Counterfeit Goods, TechCrunch (Jan. 16, 2017) https://techcrunch.com/2017/01/16/alibaba-big-data-anti-counterfeiting-alliance.
[33] Lemley, supra note 15, at 33
[34] See Lior Rokach, Introduction to Machine Learning, Slideshare 3 (July 30, 2012), https://www.slideshare.net/liorrokach/introduction-to-machine-learning-13809045.
[35] Id. at 4.
[36] Id. at 10.
[37] See Lars Marius Garshol, Introduction to Machine Learning, Slideshare 26 (May 15, 2012) https://www.slideshare.net/larsga/introduction-to-big-datamachine-learning.
[38] Id.
[39] Id.
[40] See Lei Yu et al., Dimensionality Reduction for Data Mining – Techniques, Applications and Trends, Binghamton University Computer Science 11, http://www.cs.binghamton.edu/~lyu/SDM07/DR-SDM07.pdf (last visited Feb. 23, 2018).
[41] Id.
[42] See Rokach, supra note 34, at 10.
[43] Yu et al., supra note 40.
[44] Laurens van der Maaten et al., Dimensionality Reduction: A Comparative Review, Tilburg Centre for Creative Computing, TiCC TR 2009-005, Oct. 26, 2009, at 1 (“In order to handle such real-world data adequately, its dimensionality needs to be reduced. Dimensionality reduction is the transformation of high-dimensional data into a meaningful representation of reduced dimensionality. Ideally, the reduced representation should have a dimensionality that corresponds to the intrinsic dimensionality of the data. The intrinsic dimensionality of data is the minimum number of parameters needed to account for the observed properties of the data”).
[45] Andrew Ng, Nuts and Bolts of Applying Deep Learning (Andrew Ng), YouTube (Sept. 27, 2016), https://www.youtube.com/watch?v=F1ka6a13S9I.
[46] Andrew Ng, Model Selection and Train/Validation/Test Sets, Machine Learning, https://www.coursera.org/learn/machine-learning/lecture/QGKbr/model-selection-and-train-validation-test-sets (last visited Feb. 23, 2018).
[47] Id.
[48] See Rokach, supra note 34, at 10.
[49] Alex Rampell & Vijay Pande, a16z Podcast: Data Network Effects, Andreesen Horowitz (Mar. 8, 2016), http://a16z.com/2016/03/08/data-network-effects/.
[50] James Manyika et. al., Big Data: The Next Frontier for Innovation, Competition, and Productivity, McKinsey Global Inst., May 2011, at 11, available at https://www.mckinsey.com/~/media/McKinsey/Business%20Functions/McKinsey%20Digital/Our%20Insights/Big%20data%20The%20next%20frontier%20for%20innovation/MGI_big_data_exec_summary.ashx.
[51] Id.
[52] Leonid Bershidsky, Why Are Google Employees So Disloyal?, Bloomberg (July 13, 2013, 11:41 AM), https://www.bloomberg.com/view/articles/2013-07-29/why-are-google-employees-so-disloyal-.
[53] Id.
[54] Rob Valletta, On the Move: California Employment Law and High-Tech Development, Federal Reserve Bank of S.F. (Aug. 16, 2002), http://www.frbsf.org/economic-research/publications/economic-letter/2002/august/on-the-move-california-employment-law-and-high-tech-development/#subhead1.
[55] Id.
[56] See Quentin Hardy, IBM, G.E. and Others Create Big Data Alliance, N.Y. Times (Feb. 15, 2015), https://bits.blogs.nytimes.com/2015/02/17/ibm-g-e-and-others-create-big-data-alliance.
[57] See, e.g., Finicity and Wells Fargo Ink Data Exchange Deal, Wells Fargo (Apr. 4, 2017), https://newsroom.wf.com/press-release/innovation-and-technology/finicity-and-wells-fargo-ink-data-exchange-deal.
[58] Alice Corp. Pty. Ltd. v. CLS Bank Int’l, 134 S. Ct. 2347 (2014).
[59] Karen E.C. Levy, Relational Big Data, 66 Stan. L. Rev. Online 73, 73 n.3 (2013), https://review.law.stanford.edu/wp-content/uploads/sites/3/2013/09/66_StanLRevOnline_73_Levy.pdf (explaining that the big data phenomenon is due to the need of practices to analyze data resources).
[60] Christine L. Borgman, The Conundrum of Sharing Research Data, 63 J. Am. Soc’y for Info. Sci. & Tech. 1059, 1059-60 (2012) (discussing the lack of data sharing across various industries).
[61] See Michael Mattioli, Disclosing Big Data, 99 Minn. L. Rev. 535 (2014).
[62] See id. at 545-46 (discussing the technical challenges in merging data from different sources, and issue of subjective judgments that may be infused in the data sets).
[63] See id. at 552 (discussing how institutions with industrial secrets may rely on secrecy to protect the big data they have accumulated).
[64] See id. at 570 (“[T]he fact that these practices are not self-disclosing (i.e., they cannot be easily reverse-engineered) lends them well to trade secret status, or to mere nondisclosure”).
[65] Id.
[66] Id. at 552.
[67] Patrick Clark, The World’s Top Economists Want to Work for Amazon and Facebook, Bloomberg (June 13, 2016, 10:47 AM), https://www.bloombergquint.com/technology/2016/06/09/the-world-s-top-economists-want-to-work-for-amazon-and-facebook (“If you want to be aware of what interesting questions are out there, you almost have to go and work for one of these companies”).
[68] Bill Franks, Taming the Big Data Tidal Wave 20 (2012) (discussing that the biggest challenge in big data may not be developing tools for data analysis, but rather the processes involved with preparing the data for the analysis).
[69] See Borgman, supra note 60, at 1070 (“Indeed, the greatest advantages of data sharing may be in the combination of data from multiple sources, compared or “mashed up’ in innovative ways.” (citing Declan Butler, Mashups Mix Data Into Global Service, 439 Nature 6 (2006))).
[70] Jack Clark, Apple’s Deep Learning Curve, Bʟᴏᴏᴍʙᴇʀɢ Bᴜsɪɴᴇssᴡᴇᴇᴋ, (Oct 29, 2015) https://www.bloomberg.com/news/articles/2015-10-29/apple-s-secrecy-hurts-its-ai-software-development.
[71] Kewanee Oil v. Bicron Corp., 416 U.S. 470, 490 (1974).
[72] See infra Section III-B.
[73] Brenda Simon & Ted Sichelman, Data-Generating Patents, 111 Nw. U.L. Rev. 377 (2017).
[74] Id. at 379.
[75] Id. at 414.
[76] Id. at 415 (“[B]roader rights have substantial downsides, including hindering potential downstream invention and consumer deadweight losses . . .”).
[77] Id. at 417.
[78] Rampell & Pande, supra note 49.
[79] Lina Kahn, Amazon’s Antitrust Paradox, 126 Yale L.J. 710, 785 (2017) (“Amazon’s user reviews, for example, serve as a form of network effect: the more users that have purchased and reviewed items on the platform, the more useful information other users can glean from the site”).
[80] Simon & Sichelman, supra note 73, at 410.
[81] See infra Section III-A.
[82] Id.
[83] See infra Section III-B.
[84] Jack Clark, Apple’s Deep Learning Curve, Bʟᴏᴏᴍʙᴇʀɢ Bᴜsɪɴᴇssᴡᴇᴇᴋ (Oct 29, 2015), https://www.bloomberg.com/news/articles/2015-10-29/apple-s-secrecy-hurts-its-ai-software-development.
[85] Peer, supra note 17 (“The idea that the data will be used by unspecified people, in unspecified ways, at unspecified times . . . is thought to have broad benefits”).
[86] See Harshavardhan Achrekar et al., Predicting Flu Trends using Twitter data, IEEE Conference on Comput. Commc’ns. Workshops 713 (2011), http://cse.unl.edu/~byrav/INFOCOM2011/workshops/papers/p713-achrekar.pdf.
[87] Jordan Crook, Uber Taps Foursquare’s Places Data So You Never Have to Type an Address Again, TechCrunch, (May 25, 2016) https://techcrunch.com/2016/05/25/uber-taps-foursquares-places-data-so-you-never-have-to-type-an-address-again/.
[88] See Rampell & Pande, supra note 49.
[89] See Mattioli, supra note 61, at 554 (“A final limitation on patentability possibly relevant to big data is patent law’s requirement of definiteness”).
[90] See Nautilus, Inc. v. Biosig Instruments, Inc., 134 S. Ct. 2120 (2014).
[91] See Alice Corp. Pty. Ltd. v. CLS Bank Int’l, 134 S. Ct. 2347 (2014).
[92] Robert R. Sachs, Two Years After Alice: A Survey of the Impact of a “Minor Case” (Part 1), Bilski Blog (June 16, 2016), http://www.bilskiblog.com/blog/2016/06/two-years-after-alice-a-survey-of-the-impact-of-a-minor-case.html.
[93] Id.
[94] Stephanie E. Toyos, Alice in Wonderland: Are Patent Trolls Mortally Wounded by Section 101 Uncertainty, 17 Loy. J. Pub. Int. L. 97,100 (2015).
[95] Id.
[96] Alice Corp. Pty. Ltd. v. CLS Bank Int’l, 134 S. Ct. 2347, 2349 (2014).
[97] Id.
[98] Id.
[99] Id. at 2350 (emphasis added) (citation omitted).
[100] Id. at 2357 (emphasis added) (alteration in original) (citation omitted).
[101] Id. at 2350.
[102] Id. at 2351.
[103] Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1330 (Fed. Cir. 2016).
[104] Bascom Glob. Internet Servs. v. AT&T Mobility LLC, 827 F.3d 1341, 1349 (Fed. Cir. 2016).
[105] McRO, Inc. v. Bandai Namco Games Am. Inc., 837 F.3d 1299, 1308 (Fed. Cir. 2016).
[106] Id.
[107] Bascom, 827 F.3d at 1349.
[108] Enfish, 822 F.3d at 1330.
[109] McRO, Inc., 837 F.3d at 1314.
[110] Id.
[111] Bascom, 827 F.3d at 1349.
[112] Id.
[113] Id.
[114] See Toyos, supra note 94, at 121; DDR Holdings, LLC v. Hotels.com, 773 F.3d 1245, 1257 (Fed. Cir. 2014).
[115] See supra Section II-A.

Fitting Marrakesh into a Consequentialist Copyright Framework

Fitting Marrakesh into a Consequentialist Copyright Framework
Download a pdf version of this article here. The Marrakesh Treaty to Facilitate Access to Published Works by Visually Impaired Persons and Persons with Print Disabilities entered into force on September 30, 2016. The treaty aims to alleviate what has been described as the “book famine,” and has been lauded as a significant achievement in advancing the rights of and promoting equal opportunity for the visually disabled. Contracting states are required to implement copyright limitations and exceptions to facilitate access to copyrighted material for the global print-disabled community. This note will argue that, notwithstanding the treaty’s strong rights-based underpinnings, the treaty aligns comfortably with U.S. consequentialist copyright justifications. This note will also demonstrate the limitations of other copyright justificatory theories while discussing their incompatibility with the treaty’s philosophy.

Through the Looking Glass: Photography and the Idea/Expression Dichotomy

Through the Looking Glass: Photography and the Idea/Expression Dichotomy
Download a pdf version of this article here. Copyright law has always expressed an idea/expression dichotomy, where copyright protection extends not to an idea of a work but only to work’s expression of that idea. Alas, this distinction walks a fine line with regard to non-textual and visual works. In particular, courts are prone to inconsistent outcomes and a violation of the fundamental precepts of copyright law because courts often succumb to shortcomings in grasping aesthetic theories of originality, realism, and ideas idiosyncratic to visual works. However, this dilemma may be solved within the existing framework of copyright law. This note argues that the solution should start by focusing less on visual works’ subject matter, but rather elements of the work, such as the originality and realism of the expression that clarify the author’s creativity. Moreover, the concept of an “idea” should be defined broadly as the residual locus of uncopyrightable elements in a work, rather than a cohesive concept that attempts to definitively pin down the “idea” behind that individual work. Taking this two-pronged solution would thus both recognize visual and photographic work’s unique niche within copyright as well as align these forms of art with copyright’s law’s ultimate objective of authorship protection.

Un-Blurring Substantial Similarity: Aesthetic Judgments and Romantic Authorship in Music Copyright Law

Un-Blurring Substantial Similarity: Aesthetic Judgments and Romantic Authorship in Music Copyright Law
Download a PDF version of this article here.

By Nicole Lieberman*

Introduction

By refusing to acknowledge[1] the aesthetic judgments inherent in determining copyright disputes,[2] American courts have plagued our copyright law with subjective bias[3] and doctrinal confusion. To avoid the appearance of impropriety, since at least 1903[4] courts have side-stepped clearly defining foundational concepts such as “originality,”[5] “authorship,” and “infringement.” As such, they have failed to provide a meaningful methodology for determining when a work infringes the copyright of another.[6] By instead relying on the impossibly vague “substantial similarity” test,[7] courts have crafted an impressionistic doctrine that has drifted far from copyright’s original economic purpose of incentivizing creation.

While copyright infringement requires proof of copying, mere copying is not the end of the inquiry, as “[t]rivial copying is a significant part of modern life.”[8] Thus, proof of copying, or “copying-in-fact,” is only a threshold issue for proving infringement.

Copying-in-fact can be shown through direct evidence, such as testimony, but with witnesses and honest thieves often lacking, copying is most often shown by circumstantial evidence. Indirect proof of copying is provided by evidence creating an inference that the defendant copied – typically a combination of evidence of access to the plaintiff’s work and similarities probative of copying. While courts allow expert analysis and dissection to aid them in inferring copying, the largely unguided impression of lay observers determines the more exacting question of misappropriation.[9]

Yet determining misappropriation requires an understanding of the “axiom of copyright law that the protection granted to a copyrightable work extends only to the particular expression of an idea and never to the idea itself.”[10] Application of the “idea/expression” distinction requires delicate line-drawing to decide the appropriate “‘level of abstraction’ at which one defines the ‘idea’ that merges with the subject’s expression.”[11] But fact finders are unlikely to understand on their own which “ideas” are excluded or what elements fall into the category of “ideas.” While jury instructions can theoretically work to inform jurors to exclude such elements, “in practice jurors aren’t going to know what things are, for example, scène à faire[12] in the music industry, without some testimony on standard chord progressions.”[13] Thus, jurors are not likely to understand such an ephemeral distinction between ideas and expression, especially when applied to areas in which they lack expertise, as is often the case with copyright.[14] Because the issue of misappropriation is so dependent on the interpretation of these underlying principles of copyright law, classifying the issue as purely a question of fact for the jury requires reconsideration.[15]

Courts recognize the need for expert analysis and dissection in determining infringement in cases involving computer software. Distinguishing computers as “complex” and having elements dictated by limited options, courts apply a special test to ensure only protected elements are considered for infringement purposes. Yet they proscribe such guidance when the “aesthetic arts” are at issue, failing to recognize traditions unique to genres, that all art is capable of being broken down into constituent elements, and that such elements are dictated by genre and functional constraints. Courts have assumed that art is intuitive, simply reflecting emotions, and capable of being understood by anyone. Their narrow understanding of art comes from our law’s founding.

Copyright arose in an era where the courts viewed creativity as coming from a place of pure autonomous genius,[16] but this romantic view of aesthetics is a relic of the past: a counteraction to the age of enlightenment and rationalization. The reality is that creative borrowing is almost unavoidable and results in widespread use of unprotected elements from preexisting works. Without expert guidance and the ability to dissect protectable and unprotectable elements, judges and jurors are “more likely to find infringement in dubious circumstances, because they aren’t properly educated on the difference between protectable and unprotectable elements.”[17]

Due to the prevalence of music copyright infringement suits, and the fact that music is more perceptively derivative than other media,[18] it seems disproportionately plagued by the courts’ bias for traditional aesthetics. But music, like all arts, is inherently complex and technical,[19] and few “ordinary observers” know the elements and factors that go into its creation,”[20] especially with works of less familiar genres. Thus, fact finders are easily misled into finding substantial similarity based on unprotectable elements.[21]

While music may be uniquely crippled by our current copyright regime, the problems plaguing music copyright stem directly from a lack of guidance where it is arguably most needed: the technical issue of misappropriation.[22] With fact finders less likely to detect similarities attributable to common sources in unfamiliar aesthetic, the current system results in a prejudice against lesser-known aesthetics, and a bias for the traditional. [23] The result is far from encouraging aesthetic progress.

This paper will argue that to create a more encompassing[24] and objective copyright law, that fosters progress in all arts, it is vital to expand the role of analytic dissection and expert testimony to the misappropriation prong of the infringement test.

Part I of this paper provides background on the history of court treatment of music copyright and lays out the two major approaches to copyright infringement. In addition, this part outlines the foundational principle that only the expression of an idea is protectable. Part II illustrates how the tests have veered away from the original purposes and values underlying the inquiry. It argues that by relying on the ordinary observer test for misappropriation, the tests fail to accurately account for the idea-expression distinction. In outlining the problems facing music under our current copyright regime, this section shows how the problems with the audience test are particularly problematic for music, a medium in which the line between idea and expression is often not “spontaneous and immediate” to the ordinary observer.[25] The recent “Blurred Lines”[26] lawsuit serves to illustrate how the lack of objectivity in our current law results in inconsistent application, thereby diminishing incentives to create new works. More broadly, this section considers that while the problems for music are often more noticeable than for other media, they merely expose the larger inaccuracies of the audience test. Finally, Part III considers proposals for creating a more guided and objective infringement analysis. Ultimately, this paper concludes that the best solution is adopting the test for computer software, the abstraction-filtration-comparison method (AFC),[27] as a uniform test for infringement.

Requiring careful dissection of unprotected elements by the court would ensure educated decisions, and reserving the intuitive question of whether the defendant copied those elements for the trier would preserve the economic rationale of the lay listener test.[28] Effectively reversing the analysis of proof “will likely result in greater attention to the limiting doctrines of copyright law”[29] and the evolution of reasoned rule of law.[30] By basing aesthetic nondiscrimination in objective and reasoned criteria, as opposed to the “anti-intellectual and book burning” philosophy[31] of visceral impressions, the courts can determine actual illicit copying while being receptive to unconventional aesthetics.[32]

I. Historical Background of Music in Copyright

While copyright law struggles to deal with the fine arts as a whole,[33] particular problems arise in the context of musical works. These issues are rooted in the history of copyright law. Many of the problems facing music copyright lie in the fact that creators are seeking protection under a scheme created for the distinct purpose of protecting works of literature.[34] However, these problems are not unique to music. American copyright law is based on a concept of authorship ill-suited to progress in general. This section will outline the evolution of our copyright infringement doctrine. In considering the historical application of the doctrine to musical works, this section analyzes the aesthetic norms embedded within judges’ and jurors’ findings of infringement.

A. Music’s Initial Encounters in Early Legislation and Case Law

Article 1, Section 8 of the Constitution authorizes federal legislation “[t]o promote the Progress of Science and useful Arts,”[35] but gives little guidance in defining the scope of the copyright system. The original Copyright Act of 1790 extended protection only to maps, charts, and books.[36] Though musical compositions were routinely registered under the 1790 Act as “books,”[37] it was not until the Copyright Act of 1831 that Congress expressly extended protection to musical compositions. Congress’s early failure to provide well-crafted protection for musical compositions is hardly surprising given the 1790 Act’s roots in Great Britain’s Statute of Anne, which covered only the distinct category of “books.”[38]

With no other protection available against infringers, composers naturally came to seek protection of their works through copyright.[39] Yet utilizing a scheme created for “written” works meant obtaining copyright protection solely with respect to the underlying composition, “the notated, written score, including the music and any lyrics.”[40] While seemingly analogous, music as a performing art is “often related in some way to performance” and must be understood by reference to its context, that is, elements outside the composition.[41] Though federal law since 1976 has applied copyright protection to musical recordings, including some performance elements such as percussion, recordings are treated as distinct expressions with separate copyright protection.[42] Consequently, musical compositions are protected only within the restrictive framework of the “musical work,” which is defined as a combination of melody and harmony.[43]

More problematically, courts analyzing music copyright cases tend to place undue weight on melody, rather than harmony and rhythm,[44] failing to consider the complexity of music and a realm of possible distinguishing features. Focusing on elements of music that “lend themselves to notation”[45] may seem adequate in analyzing works from European musical traditions, which typically have predominant harmonic and melodic structures,[46] but doing so fails to consider music in its totality. Because music is inherently relational,[47] our perception of musical works, and their meaning, is dependent on the context in which notes and pitches in the melody are played.[48] Elements such as timbre and spatial organization are also relevant to the way we hear music and to the similarities we perceive. Consequently, “originality is better viewed as a function of the interaction and conjunction of these elements than of any element alone.”[49]

Neglecting to consider the totality of elements in musical works, while ill-suited even to classical traditions, most drastically affects works outside of Western traditions. The main aesthetic features of non-Western music often fall outside the confines of the court’s emphasis on melody and notation. Most notably, hip-hop, which finds its roots in certain African musical traditions, features a dominant “oral tradition,” evidenced in the practice of rapping, and complex rhythmic structures with less emphasis on melodic and harmonic structures. Moreover, such traditions predominately feature the element of musical borrowing through the practice of sampling, looping, and interpolation. These features are also found in other African American musical genres, including blues, jazz, rhythm and blues, gospel, Soul, rock, reggae, funk, disco, and rap, and are found mixed with all types of music today.[50] Electronic music producers are now producing hip-hop tracks,[51] and even pop-country artists are making rhythm-centric tracks that reference hip-hop culture. Not surprisingly, entering the arena puts artists at risk of facing a copyright suit. Taylor Swift recently faced a $42 million infringement claim for using the lyric “haters gone hate,” a staple in hip-hop culture and music, [52] in her recent dance-pop track “Shake it Off.”[53] Despite the prevalence of non-notational elements, copyright’s bias for written work places works that do not fit the mold “at the bottom of the hierarchies of taste,”[54] making findings of original elements in allegedly infringing works more difficult to obtain.

Borrowing similarly conflicts with Western ideals of creativity and originality, with the result that music has historically been disvalued. Records from the time of the Statute of Anne’s enactment are telling of the hostile attitudes facing music. While literature was held in high esteem for its educative role, music was seen as an unnecessary luxury that served merely as entertainment.[55] In Pyle v. Falkener,[56] an early case brought under the Statute of Anne, defendant publishers argued that, in contrast to works of literature, authorship of music required “a high standard of originality to qualify for protection under any legal theory.”[57] Underlying their challenge was the commonly held notion that composers “merely borrowed” from “[o]ld [t]unes which had been [u]sed in [c]ommon by all persons for many years before…” and as such have no proprietary rights.[58]

Such disparaging views of music are less surprising when one considers the rise of the Romantic view of authorship in the nineteenth and twenties centuries.[59] Unlike the classical conception of authorship, which “conceives of art as imitating universal truths and ideas,” and thus contemplates the evolutionary nature of art, the Romantic view conceptualizes the creation of art “as a process that reflect[s] the emotions and personality of the individual artist.”[60] With the Romantic view informing cultural assumptions, originality often came to be defined as requiring independent creation, “which essentially appears to rule out or significantly limit borrowing.”[61]

With the functional and genre constraints inherent to music,[62] tensions existed early on in applying copyright to musical works. Yet “use of existing works has historically been a core feature of the musical composition process”[63] and the artistic process in general. The courts’ neglect to appreciate the reality of borrowing has often resulted in overbroad copyrights, extending protection to more than just the particular arrangement of the literal elements of a work.[64] Additionally, by failing to recognize distinguishing features of songs that lie outside the melody and notation, courts often find infringement based on unprotected elements. Genres that explicitly sample existing works, such as hip-hop, have been hit hardest. As a result, the courts are perpetuating a bias for traditional aesthetics at the expense of progressive and unfamiliar artistic movements.[65]

B. The Idea-Expression Distinction

Early on, courts using the idea-expression dichotomy to distinguish between unprotectable and protectable aspects of works did so on the basis of tangibility. Though the courts still use these terms in filtering out unprotectable elements, changing views on the nature of the artistic process have distorted the original tangibility basis, leading to ad hoc judicial determinations. With the rise of the Romantic view, artistic works in their entirety came to be regarded as reflecting the artist’s contributions.[66] As a result, perceptions regarding the moral (and thus, intellectual property) rights of an artist expanded to include more than just the particular arrangement of the literal elements of a work.

Originally, American copyright law viewed “ideas” as “intangible, unexpressed concept[s] that existed only in the author’s mind.”[68] Therefore, copyright protected only the tangible[69] “expression,” or the “arrangement of words which the author has selected to express the idea.”[70] The rationale served the purposes of the intellectual property clause well, since free access to ideas is critical to the development of creative works.[71] Moreover, the right granted did not include a right over certain words used, because “they are the common property of the human race.”[72] This early approach was consistent under the classical conception of the creative process, which views the artist as portraying an intangible idea or truth which “cannot and should not be captured or controlled by one artist.”[73]

However, with the rise of the Romantic view in the nineteenth and twenties centuries,[74] Congress no longer limited “expression” to the arrangement of the literal elements of the copyrighted work, but expanded it to include underlying “original” conceptual elements as well.[75] In 1909, Congress both enlarged the category of works eligible for protection and expanded the rights provided to copyright owners, including use of the work in a different medium.[76] Protection under the Act was no longer limited to the literal form or features of the expressed idea, but extended to elements of a work that are intangible and conceptual.[77] In applying the new act, the Supreme Court in Kalem Co. v. Harper Bros.[78] found the defendant’s film to infringe upon the plaintiff’s copyright in the book Ben Hur because the film expressed the same underlying idea, or plot, albeit in an entirely different medium.[79]

It became clear that balanced against the idea-expression distinctions is the countervailing consideration that copyright infringement cannot be limited to exact copying, “else a plagiarist would escape by immaterial variations.”[80] The problem is one of line-drawing: at what point is a variation distinguishable enough to “sufficiently alter a work’s substantial similarity to another so as to negate infringement,” without extending protection to the underlying idea of the plaintiff’s work?[81]

Views on the nature of art and the creative process have only continued to evolve and become more inconsistent with the idea-expression dichotomy.[82] The conceptual art movement advanced the rejection of any distinction between an artist’s idea and the ultimate expression.[83] As conceptual artist Sol LeWitt stated, “the idea or concept is the most important aspect of the work. When an artist uses a conceptual form of art, it means that all of the planning and decisions are made beforehand and the execution is a perfunctory affair. The idea becomes a machine that makes the art.”[84] In rejecting the Formalist tradition, which defined art by its form and structure, conceptual art judges art by what it contributes to the conception and definition of “art.”[85] Even an unchanged item from the grocery store, like a box of Brillo soap pads,[86] can be art if framed in a new way.

With Romantic and neo-romantic views challenging classical aesthetic theory, no universally accepted philosophical or objective basis remains for distinguishing ideas from expression in works of art. Continuing to use the terms leaves courts to make infringement decisions on the basis of their own subjective assessments of a work’s artistic value.[87] Judicial determinations of what constitutes the “idea” versus the “expression” have come to reflect personal assumptions and experiences. Courts tend to find elements of a work to be an “idea” when they are familiar with the work’s aesthetic tradition and can recognize the elements as commonplace.[88] Conversely, courts are more likely to find elements of works in less familiar traditions to be original “expression,” making them more inclined to find later uses infringing.[89] As the Ninth Circuit admitted, “‘At least in close cases, one may suspect, the classification [of idea and expression] the court selects may simply state the result reached rather than the reason for it.’”[90] Thus, with changing views on the creative process, “it is no longer necessary or valuable or even possible to dissect a work of art to uncover the universal truths or ideas which must remain freely available to all future authors.”[91]

Distinguishing between ideas and expression is perhaps most illusory in the context of music, due to the relatively limited number of compositional choices when compared with literary works.[92] Western music, at issue in most copyright suits, is primarily written in the tonal system, an organized and relational system of tones (e.g., the notes of a major or minor scale) in which one tone becomes “the central point to which the remaining tones are related.”[93] Because there are a limited number of possible pitch and harmonic relationships, options within tonal music are somewhat dictated by the system.[94] Moreover, because the tonal system is built on a hierarchy of predominate chords and pitches,[95] certain “patterns and tendencies are . . . common to virtually all musical works composed in the tonal system.”[96] The distinction between these unprotectable ideas and the original expression thereof is difficult to see, and thus the room for bias is most apparent.[97]

C. Evolution of the Copyright Infringement Tests

Courts since the nineteenth century have attempted to separate the issue of copyright infringement into two issues. First, “Copying-in-Fact”: did the defendant see and copy from the copyrighted work or did he create his work independently?; second, “Misappropriation”: did the defendant appropriate too much of the protected work?[98] The first question is used as an evidentiary tool to infer copying from access or “striking similarity,”[99] while the second focuses on the liability issue.[100] The degree of similarity between the two works is relevant to both the inquiries;[101] the phrase “probative similarity” is often used in reference to the first inquiry, while “illicit similarity” is used for the second.

Courts in the 1900s maintained the distinction between the copying-in-fact and misappropriation inquiries. A “substantial similarity” test was used for the copying-in-fact inquiry to determine whether the degree of similarity between the defendant’s and the plaintiff’s work was substantial to the point of being probative of actual copying. [102] The focus was solely on whether the defendant had copied “the labors of the original author.”[103] As such, before comparing the two works for similarities, the court filtered out unprotected elements from the plaintiff’s work, including those that were “well known and in common use.”[104]

In determining misappropriation—that is, whether the defendant copied enough of the plaintiff’s work to be held liable—courts looked to the economic or aesthetic value of what the defendant copied: if the portions extracted “embody the spirit and the force of the work…they take from it that in which its chief value consists.”[105] In this context, courts often used the adjective “substantial” to refer to the qualitative value of what was copied.[106]

As precedent evolved, courts began to combine the structure of these two prongs. As a result, courts have often confused the economic purpose of the misappropriation prong: finding infringement based solely on quantitative similarity without taking account of the unprotected elements in the original work. The analysis of the two major copyright tests below outlines how this confusion arose and focuses on the problems the misappropriation prong are causing for copyright.

1. Second Circuit Copying/Unlawful Appropriation Test

In Arnstein v. Porter,[107] the litigious Ira B. Arnstein sued the American songwriter and composer Cole Porter, alleging that many of Porter’s songs infringed the copyrights of songs written by Arnstein.[108] The Second Circuit conducted an influential bifurcated test,[109] which requires a plaintiff to prove (1) copying-in-fact and (2) illicit copying (unlawful appropriation) to establish infringement.[110]

i. Copying-in-Fact

The first prong of the Arnstein test is satisfied with the showing of (a) access; and (b) sufficient similarity, which is “probative” of copying: “The stronger the proof of similarity, the less the proof of access that is required.”[111] Thus, if similarities are so “striking” so as to “preclude the possibility that plaintiff and defendant independently arrived at the same result,” evidence of access may not be necessary.[112] Of course, the converse is not true because “access without similarity cannot create an inference of copying.” [113]

To evaluate the likelihood of copying versus independent creation, expert testimony and “analytic dissection”[114] are admissible.[115] However, the two works are to be compared in their entirety, including both protectable and non-protectable material.[116]

ii. Unlawful Appropriation

Only if the threshold issue of copying-in-fact is shown does the court move to the question of misappropriation.[117] Having established copying-in-fact, the issue of unlawful appropriation is a question of fact. Therefore, the fact finder must determine whether the taking went so far as to constitute infringement under the “substantial similarity” test.[118] That is, would the ordinary observer, unless he set out to detect the disparities, be disposed to overlook them and regard the aesthetic appeal of the two works as the same?[119] The second part of the Arnstein test is “related to the nineteenth century concern with the value of what the defendant had copied” as it asks whether the similarity “relates to material of substance and value in plaintiff’s work.”[120] However, the Arnstein test departs in some ways from earlier definitions of infringement by looking to the reaction of the “ordinary observer.”[121]

In determining misappropriation, the Arnstein test looks to “the response of the ordinary lay hearer.” That is, rather than making a purely subjective determination, the trier of fact is meant to determine the issue “in light of the impressions reasonably expected to be made upon the hypothetical ordinary observer.”[122] Because the court reasoned that the value of the work lay solely in the opinion of its intended audience, it held that expert testimony on the “impression made on the refined ears of musical experts” was “utterly immaterial.”[123] While seeming to realize the difficulty in discovering the views of the imaginary “ordinary observer,”[124] the court stated that expert testimony was permitted for the limited purpose of “assist[ing] in determining the reactions of lay auditors.”[125]

Moreover, because the court determined that the value of the works lay in their final form as impressed upon the ordinary observer, it instructed that detailed analysis and careful dissection were inappropriate.[126] Therefore, according to Arnstein, works were to be considered in their entirety, again including both protectable and non-protectable material.[127] The trier was left to depend on “some visceral reaction” as the basis for determining misappropriation.[128]

If the case involves “comprehensive non-literal similarity”[129] – that is, similarity in the overall structure of the works – the trier must make a value judgment of “whether defendant took from plaintiff’s works so much of what is pleasing to the ears of lay listeners, who comprise the audience for whom such popular music is composed, that defendant wrongfully appropriated something which belongs to the plaintiff.”[130] In a case of “fragmented literal similarity,” or verbatim copying of constituent elements, an analogous value judgment must be made, but here only with respect to the protectable portions of plaintiff’s work that have been taken.[131] Dissimilarities between materials alleged to be infringing are “significant because they mitigate any impression of similarity.”[132] Dissimilarities in other aspects of the defendant’s work, except to the extent they create an overall different impression, “typically are not significant.”[133] As Judge Learned Hand said, “no plagiarist can excuse his wrong by showing how much work he did not pirate.”[134] Thus, if the defendant copies from the plaintiff’s work, it does not matter if he adds significant material of his own,[135] resulting in what might be a transformative new work.

Consequently, under Arnstein, “[i]nstead of using some objective standards or criteria based on economic impact or quantity, courts [are] to determine infringement on an unpredictable, impressionistic basis.”[136]

iii. Further Developments and Confusions of the Arnstein Test

Although the Second Circuit in Arnstein conducted two separate inquiries into the level of similarity between the two works,[137] namely to establish copying-in-fact and then to determine misappropriation, confusion ensued from the dual use of the term “substantial similarity.” As a result of this confusion, in Ideal Toy Corp. v. Fab-Lu Ltd,[138] the Second Circuit essentially combined the issues into one subjective test. By misinterpreting the element of misappropriation identified in Arnstein as “merely an alternative way of formulating the issue of substantial similarity” rather than as an independent step, the Second Circuit stated that copyright infringement is shown solely by “substantial similarity” between the two works based on the reaction of the ordinary observer.[139] The court effectively reduced the test for prima facie copyright infringement to (1) access, and (2) misappropriation, thereby failing to consider copying-in-fact.[140] Therefore, the court rejected dissection, analysis, and expert testimony entirely and did not think it necessary to analyze the similarities between the works to determine the likelihood of independent creation.[141] By basing the entire copyright infringement inquiry on the subjective impression of those untrained in the arts, the court neglected to protect against a finding of infringement based on purely unprotectable and unoriginal elements. The Ideal Toy test fails to deal with the fundamental principle of copyright law that seeks to protect merely the expression of ideas rather than ideas themselves.[142] Unfortunately, Ideal Toy’s interpretation of the Arnstein test largely influenced the way modern courts use “substantial similarity” in determining infringement.[143]

Luckily, some courts have maintained the Arnstein two-part inquiry. In Universal Athletic Sales Co. v. Salkeld,[144] the Third Circuit restored Arnstein’s bifurcated approach in holding that a plaintiff must prove copying and “that copying went so far as to constitute improper appropriation.”[145] Moreover, the court recognized that “substantial similarity to show that the original work has been copied is not the same as substantial similarity to prove infringement.”[146]

More inherently problematic is the fact that the Salkeld court maintained Arnstein’s limits on expert analysis and dissection in determining misappropriation.[147] As a result, the Salkeldtest likewise fails to provide any objective standards or criteria for determining how much similarity is necessary to constitute misappropriation.[148]

Though courts instruct the fact finder to find misappropriation only if the defendant’s work copies not merely the idea, but ‘the expression of the idea,’ this “vague formula” is a reformulation, not a solution, to the problem of determining “what sort of similarity short of the verbatim will constitute substantial similarity.”[149] Thus, the ordinary observer continues to be left with “the impossible task of comparing only protected expression in determining substantial similarity without engaging in any thoughtful dissection or analysis of the works.”[150]

2. Ninth Circuit: “Total Concept and Feel”

The Ninth Circuit’s framework, laid out in Sid & Marty Krofft Television Prods., Inc. v. McDonald’s Corp,[151] represents the second main approach to determining copyright infringement. Although not a music infringement case, the “extrinsic/intrinsic” test developed by the Krofft court has been influential in many copyright disputes, including those involving music.[152] In recognizing that the ordinary observer is unlikely to be able to separate idea from expression in comparing two works without dissection or analysis, the Ninth Circuit proposed its own two-step test that attempts to ensure that there is “substantial similarity not only of the general ideas but of the expressions of those ideas as well.”[153] Though the test was later reformulated to include specific expressive elements during the extrinsic inquiry, as discussed in in Part I.ii.c., understanding the original formulation is key to examining its foundational flaws.

i. Extrinsic Test

The first step, or the “extrinsic” analysis, as originally cast by the Krofft court, was an objective comparison by the court for similarity in ideas.[154] Only if a substantial similarity of objective criteria under the “extrinsic” test is found do courts consider misappropriation under the “intrinsic” analysis.[155] Thus, the extrinsic test aims to limit protection to protectable elements by first filtering out unprotectable elements, including ideas, facts, and scènes à faire, and then determining whether the allegedly infringing work is “substantially similar to the protectable elements of the artist’s work.”[156]

According to the Ninth Circuit, in filtering out unprotected elements the extrinsic test incorporates the idea-expression dichotomy by limiting the scope of copyright protection to expression. As the court stated:

By creating a discrete set of standards for determining the objective similarity of literary works, the law of this circuit has implicitly recognized the distinction between situations in which idea and expression merge in representational objects and those in which the idea is distinct from the written expression of a concept….[157]

Courts conducting the extrinsic test “must take care to inquire only whether the protectable elements, standing alone, are substantially similar.”[158] Therefore, analytic dissection and expert testimony presented by the plaintiff on the similarities between the plaintiff’s work and defendant’s work are recommended.[159]

In performing the analytic dissection, courts are instructed to list and analyze the “measurable, objective elements” of the works,[160] including “the type of artwork involved, materials used, and the subject matter.”[161] For literary works, courts have listed such elements as “plot, theme, dialogue, mood, setting, pace, characters, and sequence of events.”[162] Because these factors do not readily apply to visual works of art, the court looks to the “objective details of the appearance.”[163] Without attempting to provide an exhaustive list of relevant factors, the court in Ninth Circuit listed such elements as “the subject matter, shapes, colors, materials, and arrangement of the representations.”[164]

Though described as a factual question, because it bases the question on objective criteria rather than the response of the trier, the extrinsic test may often be decided as a matter of law.[165]

ii. Intrinsic Test

Much like the Second Circuit’s ordinary observer test, the intrinsic test is entirely subjective and based on the “response of the ordinary reasonable person” to the “total concept and feel” of a work,[166] excluding expert testimony and dissection. Similar to the Arnstein court’s language of “lay [persons], who comprise the audience,”[167] the Krofft court suggested that the fact finder’s reaction be geared towards that of the intended or likely audience.[168] In a suit involving the characters in a children’s television show, the court stated, “The present case demands an even more intrinsic determination because both plaintiffs’ and defendants’ works are directed to an audience of children.”[169] Therefore, the court limited the inquiry to the understanding of a child and found substantial similarity despite noted differences. Without expert testimony to aid the trier in determining whether children might detect distinctions, the court relied on the triers’ subjective belief that children would be unlikely to notice minor distinctions.[170]

iii. Further Developments and Confusions of the Krofft Test

In Shaw v. Lindheim,[171] the Ninth Circuit modified the extrinsic/intrinsic test in recognition of the fact that district courts were not limiting the extrinsic test to a comparison of ideas.[172] Recognizing that the similarity of ideas prong is often shown by “focusing on the similarities in the objective details of the works,”[173] the Shaw court explained that the extrinsic/intrinsic test is no longer divided by an analysis of ideas and expression.[174] Rather, the extrinsic test is an objective analysis of specific “manifestations of expression,” while the intrinsic test is a subjective analysis of expression by the fact finder, which is no more than the lay observer’s visceral reaction which is “virtually devoid of analysis.”[175] Though the Shaw court recognized that the test was “more sensibly described as objective and subjective”[176] courts have confusingly continued to use the extrinsic/intrinsic language. Moreover, subsequent cases have left the analysis of improper appropriation to the jury analyzing the works as a whole.

In Swirsky v. Carey,[177] the Ninth Circuit applied the extrinsic/intrinsic test to a case involving musical works. Swirsky and his co-writer filed a copyright infringement suit claiming that Mariah Carey’s song “Thank God I Found You” plagiarized their song “One of Those Love Songs.”[178] The court rejected the district court’s approach to the extrinsic test, which involved a “measure-by-measure comparison of melodic note sequences.”[179] The Ninth Circuit felt comparing notes would fail to consider other relevant elements such as “harmonic chord progression, tempo, and key” as “it is these elements that determine what notes and pitches are heard in a song and at what point in the song they are found.”[180] precisely which elements the court should consider, explaining in dicta that the copyright framework is difficult to apply to aesthetic works such as music which are “not capable of ready classification into . . . constituent elements” the way literary work can be classified into “plot, themes, mood, setting, pace, characters, and sequence of events.”[181]

The Ninth Circuit’s opinion in Swirsky is also relevant for its proposition that substantial similarity can be found based on a combination of elements, “even if those elements are individually unprotected.”[182] For example, in Three Boys Music Corp. v. Bolton,[183] the Ninth Circuit upheld the jury’s finding that two songs were substantially similar due to the presence of the same five individually unprotectable elements: “(1) the title hook phrase (including the lyric, rhythm, and pitch); (2) the shifted cadence; (3) the instrumental figures; (4) the verse/chorus relationship; and (5) the fade ending.”[184] Even though courts filter out unprotected elements such as expression that are commonplace within a genre, they reconsider these elements in examining whether there is a unique combination of elements.[185] However, the protection granted to a unique combination of elements is “thin,” applying only to the combination itself, not the individual elements, and protecting only against “virtually identical” copying.[186]

II. The Roots, Flaws, and Legacy of Arnstein

Though Arnstein remains the majority approach to analyzing copyright infringement,[187] the opinion has attracted much criticism. This part will begin by detailing the flaws of Arnstein, and its progeny, including Krofft. Arguing that Arnstein lacks objectivity by relying on the “impression” of the lay observer and limiting the use of expert testimony,[188] this section will consider courts’ classification of misappropriation as a subjective factual question for the jury, rather than as a legal question with its own standard. This part will argue that the treatment of the “aesthetic arts” as incapable of technical analysis is the root of its subjective treatment of misappropriation. Finally, this part will argue that in relying on the general public, untrained in artistic assessment, the courts risk finding infringement based on similarities that are attributable to common sources.[189]

A. Problems with Arnstein and the Ordinary Observer

Without detailed analysis, filtering out unprotectable ideas, or guidance from experts on the artistic merits of the works at issue, little assurance remains that jurors will decide the issue of misappropriation in keeping with the law. This section will explore why relying on the reaction of laymen is problematic in the context of copyright law. While jury instructions attempt to solve the problem, this part will examine why such abstract guidance is often more confusing than helpful, as judges themselves seem baffled by the blurry line of where an idea ends and its expression begins.

While the ordinary observer test attempts to utilize the “reasonable person” standard found in other areas of the law, its application to copyright is not analogous. In areas such as tort law, the trier is capable of placing himself in the defendant’s shoes to assess the defendant’s actions.[190] In copyright, because the trier often lacks the defendant’s artistic background, the trier cannot reasonably put himself in the defendant’s shoes to consider whether he would have been constrained to copy from the plaintiff in order to achieve the given result.[191] Consequently, the trier is asked to assess the defendant’s actions as if he is an “average lay observer” reacting to whether the defendant’s work appears to have been copied from the plaintiff’s work.[192] Yet copyright is meant to protect artists from the theft of the fruits of their labor, not from the “impression” of theft.[193] While the “spontaneous and immediate” impression of theft is “important evidence,” it cannot be the end-all-be-all test.[194] Given the complexities of copyright law,[195] the ordinary observer simply is not capable of accurately detecting very real appropriation.[196]

While the ordinary observer test has logical value in protecting the artist’s interest in the potential fruits of his labor by looking to the response of the specific market from which those fruits would derive,[197] its methodology for making that determination is lacking. One problem is that it is not clear the trier has knowledge of what constitutes the “lay listener’s” response,[198] especially considering the multitude of possible reactions even among a “target” audience.[199] While expert testimony is permitted to inform the fact finder on the views of the target audience, it is questionable whether qualifying as a music expert establishes “an expertise in the aural perceptions of a lay hearer.”[200] Put another way, “whether an expert, highly educated in the field of music theory, analysis, and history, can in fact hear again as a lay listener is speculative at best.”[201]

Prohibiting expert dissection and analysis on the ultimate issue of misappropriation deprives the trier of information that may be helpful in hearing the music through the ears of the “audience for whom such popular music is composed.”[202] As Professor Nimmer pondered:

If what is to be protected is literary theft, and not the impression of literary theft per se, why, we may wonder, must the view be “uncritical,” and why must there be no suggestion and pointing to similarity, if that suggestion would prove helpful to the trier in seeing that all or a part of plaintiff’s work formed the basis for all or a part of defendant’s work?[203]

Experts could note similarities dictated by the particular type of work at issue that are “most likely insignificant to the ears of the targeted audience familiar with that form or type of work,”[204] thereby helping to ensure that infringement is not found based on common sources or coincidence,[205] and conversely, ensuring that very real appropriation does not go undetected.[206] As Justice Clark, dissenting in Arnstein, noted, the jury is not “‘pre-eminently fitted to decide questions of musical values,’ which are different from an ordinary fact-finding exercise.”[207]

Perhaps the outcome in Arnstein, and thus the bases of our modern law, can be explained by the views of Justice Frank, who wrote for the majority. Frank believed that some judicial decisional processes “like the artistic process, involve[d] feelings that words cannot ensnare” since they contain “overtones inexpressible in words.”[208] For Frank, music was the prime example “of a hunch that was intrinsically incapable of disaggregation.”[209] He wrote, “a melody does not result from the summation of its parts; thus to analyze a melody is to destroy it. It is a basic, primary unit.”[210] Similar language is found in opinions today, including the Ninth Circuit case Swirsky cited earlier.

Attempting to break down what was underlying Frank’s opinion, one commentator argued that the Arnstein court’s exclusion of expert testimony and dissection seemed rooted in the court’s fact-skepticism and unwillingness to entrust the explanation of similarities to experts, preferring instead a “subjective determination,” perhaps in the hopes that a lay-jury trial would confirm the judge’s own subjective hunches.[211]

While courts typically rely on lay jurors to apply the law to the facts of a case at hand,[212] it seems naïve to believe jurors are capable of understanding the complexities of copyright law, particularly the ever elusive idea-expression distinction.[213] Federal judges themselves have found the doctrine difficult to apply,[214] and for good reason as “precision in marking the boundary between the unprotected idea and the protected expression . . . is rarely possible.”[215] Determining what is an “idea” versus an “expression” requires more than mere application of the law; it requires interpretation of the law and consideration of policy. Even with a basic understanding of copyright law, applying the doctrine to music is still more difficult. “What of a song’s music is ‘idea’ and what is ‘expression’?”[216]

While jury instructions attempt to inform jurors of the law,[217] in practice instructions are often an inconsistent, “confusing welter of legal jargon” that may wrongly suggest that any copying, including copying of an idea, counts as infringement.[218] Moreover, clarity in keeping the two different “substantial similarity” inquiries distinct, both in purpose[219] and procedure, is often lacking. Because both the Arnstein and Krofft tests forbid expert testimony on the ultimate issue of misappropriation, fact finders will often have to consider expert analysis and dissection on alleged similarities during the copying-in-fact or extrinsic inquiries, but somehow disregard such testimony on the misappropriation or intrinsic inquiry.[220] “Especially in complex cases, [it is] doubt[ful] that the ‘forgetting’ can be effective when the expert testimony is essential to even the most fundamental understanding of the objects in question.”[221] Moreover, courts often fail to clearly explain the two-step inquiry and the need to disregard testimony presented on the first issue in determining the second.

For example, consider a well-known music plagiarism case, Selle v. Gibb,[222] where the plaintiff, a part-time musician and composer, sued the Bee Gees, alleging that the Bee Gees’ “How Deep is Your Love” infringed the copyright of his song, “Let It End.”[223] Without evidence of access, Selle sought to establish copying-in-fact by showing substantial similarity between the two songs, relying heavily on expert testimony.[224] Yet in making the misappropriation determination, the jury was neither instructed to disregard the expert’s testimony on substantial similarity, nor informed of the two distinct steps of the Arnstein test.[225] Instead, the jury instructions combined the copying-in-fact and misappropriation steps, explaining that “[t]o prove substantial similarity plaintiff must establish…that the average person would recognize [defendant’s song] as having been appropriated from parts of [plaintiff’s song].”[226]

Consider also the jury instructions in the recent “Blurred Lines” trial. Judge Kronstadt’s first instruction to the jury on the copyright infringement standard said that any copying of original elements is unlawful:

Anyone who copies original elements of a copyrighted work during the term of the copyright without the owner’s permission infringes the copyright….[If] 1. the Gaye Parties are the owner of a valid copyright; and 2. the Thicke Parties copied original elements from the copyright work….your verdict should be for the Gaye Parties.[227]

According to this instruction alone,[228] whether the copied element is an unprotectable idea is irrelevant. In attempting to explain the extrinsic/intrinsic analysis, Kronstadt further suggested that copying an idea can count as infringement. Kronstadt instructed the jury to consider similarities in ideas, as well as expression, on the extrinsic test, but he failed to tell the jurors to disregard any similarity in ideas when considering the “concept and feel” during the misappropriation inquiry.[229] Following this questionable guidance seems to be the exact mistake the jury made in finding that Robin Thicke and Pharrell Williams infringed the copyright in Marvin Gaye’s song “Got to Give it Up” by creating their stylistically similar song, “Blurred Lines.”[230]

Fundamentally, the problem with the audience test is that by wholeheartedly relying on the lay juror, the test erroneously treats the question of misappropriation as a pure question of fact. Infringement is far less intuitive, and more complex, than ordinary negligence. Just as the testimony of medical experts is necessary in negligence cases in the context of medical malpractice, copyright is dependent on a technical analysis of works which jurors know little about. Determinations of infringement necessarily involve ad hoc line-drawing that affects artistic incentives and the public’s access to art works. While it is possible to limit and tailor the test, as it stands, it has been disappointingly inaccurate and has been often used as a “verbal formula to explain results otherwise reached.”[231]

B. Krofft: Reconciling Arnstein and the Idea-Expression Dichotomy?

While the court in Krofft seemed to recognize the problems with the Arnstein test in its failure to ensure copyright infringement was only found on a similarity of expression, the Ninth Circuit neglected to explain how its own test would resolve any of the problems presented by the Second Circuit approach.[232] In laying out vague criteria for defining ideas versus expression, the Ninth Circuit left lower courts and artists to their own devices in figuring out where to draw the line.[233] The court itself noted that the extrinsic test is “turbid waters.” “Nevertheless,” it continued, “the test is our law and we must apply it.”[234] Applying the test to artistic works is even more problematic. As the court said in Swirsky, “the extrinsic test provides an awkward framework to apply to copyrighted works like music or art objects, which lack distinct elements of idea and expression.”[235]

By failing to state whether a lack of substantial similarity in ideas is relevant for copying or misappropriation, or both,[236] the extrinsic/intrinsic test fails to isolate the issue of copying from the issue of misappropriation. Focusing on the idea-expression distinction improperly frames the question. Similarity of ideas found during the extrinsic inquiry may be probative of copying, but such similarity does not prove copying of protected expression. Furthermore, the Ninth Circuit never made clear how the ordinary observer could determine if there is a similarity of expression while refraining from dissection and analysis. If anything, fact finders in the Ninth Circuit are left with less expert guidance than in the Second Circuit, as the Second Circuit at least allowed testimony to inform the trier of the intended audience’s likely views rather than leaving them guessing.

As in other jurisdictions using the term, courts applying Krofft struggle to define “substantial.” The focus ranges from audience confusion,[237] to the substantiality relative to the overall work,[238] to the aesthetic or financial value of the portion of the work copied.[239] In the Krofft decision itself, because the works at issue were directed at children, the court focused on the impact of the “total concept and feel” of the works “upon the minds and imaginations of young people.”[240] Thus, some courts adopting the extrinsic/intrinsic framework utilize an “intended audience” test for works that appeal to an audience with “specialized expertise,”[241] therefore allowing the use of expert testimony. Yet such works have been narrowly construed and usually only include computer software.[242] Most courts applying the Ninth Circuit test determine the likelihood of audience confusion by simply comparing the “total concept and feel” of the works to the ordinary observer as determined by jurors with no specialized training or expertise.[243]

Courts that look to the value of the elements at issue criticize the “total concept and feel” approach for failing to maintain the idea-expression distinction during the misappropriation analysis. As the Ninth Circuit stated in Cooling Systems & Flexibles, Inc. v. Stuart Radiator,[244] “What is important is not whether there is substantial similarity in the total concept and feel of the works . . . but whether the . . . amount of protectable expression in Cooling Systems’ catalog is substantially similar to the equivalent portions of [the defendant’s] catalog.”[245] The court in Cooling Systems stated that the intrinsic test must take into consideration the nature of the elements that were allegedly infringed,[246] reasoning that “the fewer the methods of expressing an idea, the more the allegedly infringing work must resemble the copyrighted work in order to establish substantial similarity. . . .”[247] Yet the court continued to prohibit the aid of dissection, analysis, and expert testimony, failing to explain how an ordinary observer lacking expertise could make such a determination.[248]

C. The Modern Music Dilemma

While creators of musical works have struggled to extend copyright beyond its natural borders since the Statute of Anne was enacted, the inexact fit of music in our copyright doctrine has only become more obvious with the invention and popularity of sampling technology. Since the advent of digital sampling technology in the 1960s, the art of taking a portion, or sample, of a sound recording and repurposing it to make a different song has become engrained in nearly every popular musical genre.[249] Yet, after the industry-shattering decisions in Grand Upright Music, Ltd. v. Warner Brothers Records Inc.,[250] and Bridgeport Music, Inc. v. Dimension Films,[251] which resulted in a bright-line rule against de minimis copying of even three notes of a sound recording,[252] artists need to seek licensing rights to sample a sound recording. However, the costs to do so are often prohibitive, both in terms of negotiating fees and contacting often-elusive copyright owners.[253] Consequently, many artists have turned to re-creating segments of prior musical compositions for use in their own works, a practice known, at least in the hip-hop industry, as “interpolation.”[254] Though creative borrowing of this type is deeply embedded in art history, after the Blurred Lines verdict it seems songs that interpolate prior works are also at risk of extinction.

While we may be lacking strong evidence that copyright actually encourages creativity,[255] copyright can suppress the creation of works of a specific nature.[256] After Grand Upright put “the fear of God” in recording companies, artists releasing a record on a major label were forced to clear every sample used.[257] As a result, songs composed of various samples from multiple sources were no longer possible. Records like Paul’s Boutique by the Beastie Boys, which is almost entirely composed of 104 samples, would be “financially and bureaucratically impossible” today.[258] Each sample would have to be cleared by obtaining two different licenses from two different rights holders: the owner of the sound recording and the owner of the underlying composition. To make matters more complicated, additional licenses are needed if the sampled song contains samples itself, as is increasingly the case today.

While artists can obtain a compulsory license to create a cover of a prior song, using only a portion of a song while substantially altering “the melody or fundamental character of a work” falls outside of the range of the statutory license,[259] even though doing so entails more originality and less substantial similarity.[260] Newcomers are back in the position of having to seek permission and pay whatever rights holders demand,[261] or face the risk of hefty legal judgments and court costs.

For well-known songs, the costs to license are often as high as 100% of the royalties generated by the new song, and sometimes higher. Mark Ronson, a music producer and sampling guru, analogizes the process of creating his debut album, Here Comes the Fuzz, to the process of producing the fictional musical in the play, The Producers. In creating the track, “Ooh Wee,” Ronson wanted to sample two different songs by two different artists: a drum sample from Dennis Coffey’s song “Scorpio,” and a string sample from a cover of Boney M’s song “Sunny.” Collectively, the owners of the rights wanted 125% of the song’s royalties. Believing the samples to be necessary to the “emotional effect” and “toughness of the beat,” Ronson said he had to do it: “I had to pro-rate my entire album down so I could rob this song to pay that guy….”[262] In hindsight, the decision may have made his career, as the song ended up being the lead single and a chart-topper.[263]

Rights holders can always refuse clearance entirely, as they often do. When musician and producer Danger Mouse put out The Grey Album, which splices together samples from The White Album[264] by the Beatles with an a cappella version of The Black Album by rapper Jay-Z, the recording company that holds the Beatles’ copyright was able to prevent the album’s distribution, despite approval from both Jay-Z and the surviving members of the Beatles.[265]

Since Grand Upright, it has become largely impossible to create songs using more than one or two samples. As a result, industry practice and the sound of hip-hop music abruptly changed. Artists turned to heavily interpolating a few samples per song, particularly from artists who are amenable to having their music sampled. Dr. Dre’s production style, which was highly influential in spawning an entire era in hip-hop, changed from outright sampling to heavily interpolating the 1970s funk band, Parliament.[266] Dre’s style changed hip-hop forever, as G-funk, defined by Parliament’s influence, became the most popular genre in hip-hop during the 1990s.[267]

While some artists are still producing heavily-sampled albums, it seems only those with the most obvious fair use defense are confident enough to do so without a license, perhaps recognizing the strength of their defense and realizing record companies would rather not risk setting bad precedent.[268] The D.J. Gregg Gillis, better known as Girl Talk, is perhaps the most notorious sampler; he uses hundreds of small samples on a single album, never licenses anything, generates tons of publicity, and is never sued. As Gillis put it, with so many samples, “[i]t would take you hundreds of hours of work and hundreds of thousands of dollars to clear the rights to this album even if you wanted to.”[269] Yet, Girl Talk’s business is not without harm. Both iTunes and a CD distributer refused to carry his most recent album, Night Ripper, because of legal concerns.[270]

While Girl Talk might feel confident in his defense, anyone using a sample in a more qualitatively and quantitatively significant way is likely to fear settlement fees, or worse, high statutory damages. Even with a seemingly good defense, those with less fame than Girl Talk, or Pharrell Williams, are less likely to take the risk. When part of the beauty of digital sampling technology lies in its removal of bars to entry, allowing a twenty-year-old kid to create a critically acclaimed album with cheap technology,[271] cases like Grand Upright and “Blurred Lines” leave the next Dr. Dre on the margins.

Although Dr. Dre’s debut album immortalized Parliament’s style as an entire hip-hop genre in and of itself, and the release of “Blurred Lines” landed “Got to Give it Up” back on the Billboard 200 after a decade’s long hiatus,[272] the courts have ignored the economic reality of homage and have placed it on par with theft. A justified reason for deeming musical borrowing as theft is lacking when artists of all mediums, from classic literature to appropriations art, have borrowed from their predecessors without anyone taking notice, or under the defense of “fair use.”[273] One rationale is that our copyright law is based in the romantic conception of authorship, that is, the notion that the author has the right to control the context of his work and should be able to object to “her work [being] sampled and added into a work she finds repugnant.”[274] However, applying moral rights arguments to the issue does not serve the Intellectual Property Clause’s purpose of striking a balance between economically incentivizing the production of creative works and maintaining public access to those works.[275]

The current tests do not adequately balance the impact of copyright on creativity. Rather, courts focus more on the potential harm to the plaintiff’s market[276] than on assuring “access to the raw materials that artists need to create in the first place.”[277] Progress and culture depend on the accumulation of aesthetic works. By granting property rights in creative building blocks and requiring new works to be wholly different from prior works, the courts are treating the arts like science and technology, where progress means improving. However, progress in the arts is valued for reasons beyond efficiency. Society benefits from the accumulation of artistic works and the mere experience of making them.[278] “Culture . . . is a social phenomenon. It is not the creation of one or another artist, but of many doing somewhat similar things.”[279] The scènes à faire doctrine is one way of protecting free access to artistic materials. But with jurors unable to appreciate the balance of these competing interests, we risk skewing it in a way that stifles creativity.

1. Scènes à Faire in Music: “Blurred Lines” as a Case in Point

Credited as a Billboard’s Song of the Summer for 2013,[280] and the best-selling single of 2013,[281] it is no wonder the controversies over Robin Thicke and Pharrell Williams’s song “Blurred Lines” made so many headlines. The song was controversial from its release date, featuring arguably misogynistic, “rapey” lyrics,[282] and a music video that was removed from, and then censored on, YouTube for violating the site’s policies regarding nudity.[283] Then came the copyright controversy. Marvin Gaye’s family accused Thicke and Williams of copying the “feel” and “sound” of Gaye’s “Got to Give It Up,” [284] the song Thicke even publicly noted as the influence for “Blurred Lines.”[285] Therefore, the issue in the “Blurred Lines” case revolves around whether Thicke and Williams illegally interpolated original elements of Marvin Gaye’s musical composition.

In response to the allegations, Thicke and Williams sought a declaratory judgment that “Blurred Lines” did not infringe upon “Got to Give It Up.”[286] When that failed, and inevitably backfired with the Gaye family filling a cross-complaint, the plaintiffs moved for summary judgment, claiming that the eight alleged similarities between the songs are based on elements of Gaye’s sound recording that are not part of the copyright that Gaye’s family claims to own.[287] According to the Gaye family’s expert report, these eight“substantially similar features” include: (1) the signature phrase; (2) hooks; (3) hooks with backup vocals; (4) the core theme in “Blurred Lines” and the backup hook in “Got to Give It Up”; (5) backup hooks; (6) bass melodies; (7) keyboard parts; and (8) unusual percussion choices.[288] The family’s expert also alleged that the songs share similar “departures from convention such as the unusual cowbell instrumentation, omission of guitar and use of male falsetto.”[289]

Though most of these elements are not part of the underlying composition, and the melodies of the two songs are arguably completely different (“one’s minor and one’s major. And not even in the same key”[290]), the court found that the Gaye family made a sufficient showing of substantial similarity to satisfy the extrinsic test and the issue went to trial.[291] The jury found against Thicke and Williams with a judgment topping $7.3 million, one of the largest ever in music copyright history.[292] While Thicke’s and Williams’s fortune and fame seem unlikely to be tarnished, with each of them earning over $5 million for the song itself, [293] the verdict shows how misguided results can be under our infringement tests. In misapplying the already confusing Ninth Circuit test for substantial similarity, the court protected musical style and genre under the guise of protecting an original combination of elements. As one composer and producer put it, the court “made it illegal to reference previous material. . . . I’m never going to come up with something so radically different that it doesn’t contain references to something else.”[294] While the creators of “Blurred Lines” have already decided to appeal the verdict,[295] if it is upheld, artists on the margin, who are less willing to take on legal fees and whose work is less in tune with romantic authorship, will be discouraged from “creating any new songs that evoke the feel of the music that inspired them in their youth.” Further, “with the length of copyright we have these days, artists who want to feel confident that their musical influences are in the public domain are going to have to go all the way back to ragtime.”[296]

The verdict reached is procedurally problematic for a few reasons. Under the current Ninth Circuit test, the court erred in allowing the jury to hear excerpts of the sound recordings for “Got to Give It Up” and “Blurred Lines,”[297] which contained elements that are not part of Gaye’s written composition and, therefore, not part of his copyright.[298] Though the background chatter, party noise, and percussion, are common to both songs and contribute to the instinctive feeling that “Blurred Lines” “reminds” us of “Got to Give It Up,” they are totally irrelevant to the issue of substantial similarity.[299] Yet, these elements seemed to sway the jury in reaching its verdict.

Many, if not all, of the elements in “Got to Give It Up” are unoriginal staples in funk music, from the walking down funky bass line to the falsetto and melisma hook,[300] but Judge Kronstadt, applying the extrinsic test on summary judgment, held that a genuine issue of material fact existed either on the similarity to protectable elements, or on the similarity to an original combination of elements.[301] To show that the allegedly infringing elements are unprotectable scènes à faire, the defendants’ expert cited multiple prior songs that used many of the same allegedly infringing elements.[302] Judge Kronstadt still ruled that the testimony was not sufficient, citing Swirsky for the idea that the defendants failed to show that the plaintiff’s work is “more similar” to prior works than it is to the defendants’ work.[303] Yet the issue in Swirsky, and here, was whether the individual elements are scènes à faire, not whether the works as a whole are unoriginal.[304] Thus, Kronstadt’s reliance on Swirsky’s “more similar to prior works” language is misplaced, and bypasses actual consideration of the protectability of the elements themselves.[305] The court thereby failed to dispose of the issue of whether the defendants copied original elements on summary judgment, leaving the question to the jury.

In erroneously applying the “more similar” to prior works test to the works as a whole, Kronstadt further failed to consider the actual test for copying a combination of unprotectable elements: whether the works are “virtually identical.” As Pharrell and Thicke argued, the requirement for originality in a combination of unoriginal elements is much more stringent than the usual “some minimal level of creativity.”[306] The elements need to be “numerous enough and their selection and arrangement original enough that their combination constitutes an original work of authorship.”[307] Furthermore, infringement can only be found if the defendant’s work is “virtually identical” to the copyrighted work.[308] While many might say the two songs sound similar, with entirely different melodies and lyrics, one cannot maintain that the songs are “virtually identical,” especially when the misappropriation inquiry is supposed to be based solely on the written melodies, chords, and lyrics.[309] Even still, copying that is virtually identical may fall within the merger doctrine – that is, it may be an unavoidable product of using the same idea, which as discussed above is not a protectable interest.[310]

Because Judge Kronstadt determined that Pharrell and Thicke failed to meet the stifling burden of “more similar” to prior works on summary judgment, the final decision went to the jury to be based on their spontaneous reaction unguided by musicologists and uninstructed by the court on the “virtually identical” standard. Thus, the simple instinct that “Blurred Lines” felt like “Got To Give It Up” determined the verdict, regardless of the law. Yet, if the “feel of a work is sacrosanct,” many songs would be illegal for simply being a part of a “derivative and rigid genre,” as is often the case in music.[311]

When genres and subgenres have been inspired by one song, privatizing common elements can only do harm. Musicians like Marvin Gaye still have the incentive to create, especially with the notoriety and revitalized publicity that comes with a new artist referencing an old song. A reference from today’s stars can make a music legend. Snoop Dogg’s cover of Slick Rick’s 1985 hit “La Di Da Di,” and Notorious B.I.G.’s sample of it in “Hypnotize Me,” made Slick Rick’s song the most sampled rap song of all time. Dr. Dre’s use of the drum-break from The Wintons’ “Amen, Brother” started a narrative that led this little known band to have arguably the most sampled song of all time. The Amen Break has been used in over 1,700 songs and has become the basis for drum-and-bass and jungle music.[312]

Despite the notoriety and financial benefits of being sampled, by looking to our skewed notions of authorship, both judges and juries alike are trigger-happy to find copyright infringement when there is so much as a waft of homage. Their intentions are well wrought in a desire to protect the “genius” of artists loved by generations. However, that desire is based on notions of fairness, and copyright was never meant to protect artists’ moral rights in their works. Incentives to create have always been the only concern, balanced against maintaining public access to creative works. Yet “discussions of copyright and its goals frequently conflate the compensatory and control aspects of copyright on the incentive side.”[313]

If we expand the copyright in an original selection and arrangement to find infringement when any song includes just some of its elements, the blurry standard determining how many elements count as copying will result in risk-averse content creators. When nothing is truly original or avant-garde in music, the safe route in creating content is not clearly laid out either. Even if artists can be completely progressive, it is not clear they will be rewarded for doing so. Popular content is popular for a reason, and the benefits of creating it will be limited to the lucky first comers who hold a monopoly on elements they did not even create simply because they strung them together first. The success of “Blurred Lines” might be due, at least in part, to its ability to incorporate so many nostalgic funk clichés at once, all in the context of a modern pop song. Moreover, the enjoyment derived by the public from having an expanse of options to choose from will be lost. Of course, the effects of prohibiting “inspired works” will touch on industries beyond music, covering the fine arts, dance, and anything covered by copyright.

III. Considering Dissection and Reversed Questions of Law and Fact

When copyright infringement is meant to prohibit the copying of protectable elements of a work, it seems that the audience test, unguided as it is, can “play no useful role” in fulfilling the goals of copyright law.[314] Expert testimony and analytic dissection are necessary to maintain the distinction, both at the copying and misappropriation stages of the inquiry. More specifically, courts should be informed fully of the broader contexts within which specific artistic works are created.

While some courts recognize Arnstein’s limits and allow expert testimony when works are of a “highly technical nature,” thus far only computer software has met this characterization.[315]

The courts’ unwillingness to see “aesthetic arts, such as music, visual works or literature” as sufficiently complex to warrant the aid of expert guidance is merely an artifact of indoctrinated opinions on the nature of aesthetics. Yet with culture becoming ever more aware of its recombinant nature,[316] and the arts becoming increasingly technical in the age of digitalization, it is necessary to alter the current infringement tests in order to encompass varying artistic theories.

Courts need to act as gatekeepers, preventing an onslaught of needless and threatening litigation by deciding whether there is a cognizable claim of misappropriation and identifying the unprotectable elements in a work before sending the intuitive issue of copying to the trier. There have been many scholarly suggestions for reforming the two-part test. This part considers adoption of a proposal first suggested by Nimmer and expanded by Professor Lemley: extending the “abstraction-filtration-comparison” (AFC) test for computer software to all cases deciding copyright infringement.[317]

First adopted by the Second Circuit in Computer Associates v. Altai,[318] the AFC test requires the court to identify which aspects of the program constitute its expression versus ideas (“abstraction”), remove from consideration unprotectable ideas (“filtration”), and only then compare whether the defendant copied the protectable elements (“comparison”).[319]

Following Altai, the AFC test was widely adopted in determining substantial similarity in the non-literal aspects of computer programs. However, as Nimmer notes, “there is no reason to limit it to that realm.”[320] In fact, the AFC method is more consistent with the Supreme Court’s definition of infringement (along with ownership of a valid copyright) as the “copying of constituent elements of the work that are original.”[321] Whether expert testimony is permitted, and to what extent, is left to the discretion of the district court.[322] This approach allows the court flexibility in determining the necessity of expert testimony depending on the complexity and nature of the works at issue. A court might therefore determine that the issue of substantial similarity in a case involving generic pop songs would best be determined by lay jurors, but in most cases where songs mix genres and traditions, the court would need the testimony of experts well versed in the nature of the art at issue.

Given that nearly every circuit already permits expert testimony on both prongs of the copyright infringement test in software cases,[323] adopting the AFC test wholeheartedly seems like a more feasible path to altering the current framework than other suggested reforms.[324] In fact, the Tenth Circuit already applies the AFC test for all copyright infringement cases, and the Sixth Circuit uses a variation of the Tenth Circuit test.[325] As discussed above, expanding the role of analytic dissection and expert testimony will result in better maintenance of the idea-expression distinction and respect for the limits to copyright protection. Moreover, allowing expert testimony in guiding the jury will solve the problem of juries having trouble disregarding the testimony they hear on the copying-in-fact prong in deciding misappropriation.

Professor Lemley suggests a further change: reserving the issue of misappropriation to the court as a question of law, to be determined on summary judgment. The role of the jury would be preserved in leaving the intuitive issue of whether the defendant copied as a question of fact.[326] According to Lemley, this seems to be the “practical import” of the AFC test.[327] Reversing the judge-jury role would better serve the interests of copyright, as the issue of which elements are protectable and unprotectable in any given work implicates substantial policy considerations and interpretation of the law better suited for a judge well-versed in the law.

Moreover, by providing written judicial opinions, this altered framework would allow a reasoned jurisprudence to develop on the issue of which elements are protected. The transparency and predictability of the law would allow artists a safe harbor and guaranteed protections in creating new works, thereby preventing the chilling effects of a potentially devastating lawsuit.

Taking reform a step further, Professor Balganesh suggests reversing the order of the test.[328] That is, have the court determine misappropriation as a question of law on summary judgment, filter out what is sent to the jury, and then have the jury determine the question of copying-in-fact of “protected expression.” The jury would still hear evidence on the issue of “protected expression,” including expert testimony on whether elements are scènes à faire or original in the area or genre, and even on the ultimate issue of similarity between the works and probative similarity. The jury would also hear the judge’s own dissection from the first step.[329] Balganesh argues that reversing the framework would encourage disposition on summary judgment, providing judges with a real gate-keeping role.[330] Moreover, it removes the subjectivity of the audience test, which he argues was based in mere skepticism towards rules, judges, and the law.[331]

While reversing the ordering might offer the benefit of preventing the judge’s determination of misappropriation from improperly influencing the jury’s probative similarity analysis, almost every court has adopted some version of the Arnstein-Krofft test,[332] which might make it difficult to persuade courts to do exactly the opposite.

At least preliminarily, extending the AFC approach seems more feasible and the need for dissection and expert testimony most ripe. Allowing the judge to determine the question of misappropriation as a matter of law, while preserving the issue of copying-in-fact for the jury, removes the inherent subjectivity and messiness of the audience test while complying with the constitutional mandate of jury trials in copyright lawsuits.[333] Furthermore, it uses juries for the role they are best suited for: the intuitive and fact-dependent question of whether the defendant copied.

On both questions, expert testimony can aid the court in understanding the nature of the work at issue. The controversy over sampling, for example, focuses on the taking rather than the contribution. Yet, experts in musicology know that sampling can be transformative. As D.J. Shadow put it, sampling is a way to reintroduce a person’s music to people “in a completely different context than the way they originally intended.”[334] Mark Ronson similarly explained that through sampling artists insert themselves in a narrative and push that narrative forward.[335]

An expert might testify that sampling extends an African American oral tradition known as “signifying,” a type of word play that draws attention to the cultural significance of the words.[336] For example, an expert could explain that the Beastie Boys’ line “I shot a man in Brooklyn just to watch him die” places Johnny Cash’s line “I shot a man in Reno just to watch him die” in a new setting – a hip-hop anthem called “Hello Brooklyn” – thereby removing the ideas of consequences and regret found in the context of Cash’s murder ballad.[337] An expert could note that while sampling might intuitively seem like theft, it is no more so theft than quoting a passage from another book in a novel.[338] Rather, sampling is a form of textual revision and a literary device through which text speaks to text.[339] Through sampling, 2 Live Crew relied on repetition of the key elements of “Oh, Pretty Woman” to achieve a parody of the song.[340]

Our current framework for copyright infringement fails to inform courts of the artistic merits behind our most illustrative and progressive artistic movements. In perpetuating the courts’ biases, copyright is its own worst enemy, stifling innovation where it is most likely to happen: on the margins.

Conclusion

Though the problems with copyright pose unique problems for music, these problems reflect the larger difficulties in our current copyright law. Our infringement doctrine inhibits progress by making “substantial similarity” the end-all-be-all test. The generalized focus on substantial similarity leaves homages and similarities inherent in the genre subject to findings of infringement because the law, like our culture, sees derivative works as unworthy. But what is originality? As music producer and sampling guru Mark Ronson said in an interview with NPR: “Well, what’s the T.S. Eliot quote, which apparently he even stole from Picasso, about ‘Genius steals…?’ ‘Good artists borrow, great artists steal.’”[341] How we define “Progress of Science and useful Arts” is rooted in the dominant cultural beliefs of aesthetic value, but innovation comes with embracing progressive ideas and more of them.[342] The law needs to protect a new kind of originality, one that might not fit the Romantic mold, but that re-conceptualizes and re-frames preexisting works and provides listeners with a different experience than the original. The most important question should be what the newcomer added: how they took influences to make something new, because progress never involves creating something from nothing.



* J.D. Class of 2016, New York University School of Law; B.A. Philosophy, Politics, and Law, magna cum laude, Binghamton University. The author would like to thank Professor Christopher Sprigman for his inspiration and guidance and the 2016-2017 Editorial Board of the NYU Journal of Intellectual Property and Entertainment Law for their encouragement.

[1] See Bleistein v. Donaldson Lithographing Co., 188 U.S. 239, 251 (1903) (“It would be a dangerous undertaking for persons trained only to the law to constitute themselves final judges of the worth of pictorial illustrations, outside of the narrowest and most obvious limits.”).

[2] See generally Alfred C. Yen, Copyright Opinions and Aesthetic Theory, 71 S. Cal. L. Rev. 247, 251, 285–86 (1998) (arguing that “judges necessarily make decisions of aesthetic significance in copyright” as implicit in determining whether elements are “public domain,” or if they are “protected material.” Moreover, substantiality rests on “how sensitive” courts are to the degree of quantitative and qualitative similarity between two works).

[3] Id. at 251 (“[S]ince no aesthetic perspective can be neutral and all-encompassing, aesthetic bias becomes inherent in copyright decisionmaking . . . .”).

[4] See, e.g., Nichols v. Universal Pictures Corp., 45 F.2d 119, 121 (2d Cir. 1930); Brandir Int’l, Inc. v. Cascade Pac. Lumber Co., 834 F.2d 1142 (2d Cir. 1987); Carol Barnhart Inc. v. Economy Cover Corp., 773 F.2d 411, 415–18, 420 (2d Cir. 1985).

[5] See Olufunmilayo B. Arewa, From J.C. Bach to Hip Hop: Musical Borrowing, Copyright and Cultural Context, 84 N.C. L. Rev. 547, 565(2006) (“Although originality is not explicitly included in the Intellectual Property Clause of the U.S. Constitution, it is a fundamental assumption of current copyright law that originality is implicitly mandated by the Constitution’s references to ‘authors’ and their ‘writings.’”) (citations omitted); see also U.S. Const. art. I, § 8, cl. 8 (“The Congress shall have Power . . . To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.”).

[6] See generally Amy B. Cohen, Masking Copyright Decisionmaking: The Meaninglessness of Substantial Similarity, 20 U.C. Davis L. Rev. 719, 720 (1987).

[7] H.R. Rep. no. 94-1476, at 61-62 (1976); S. Rep. No. 94-473, at 71 (1975) (“[A] copyrighted work would be infringed by reproducing it in whole or in any substantial part, and by duplicating it exactly or by imitation or simulation.”) (emphasis added).

[8] Davis v. Gap, Inc., 246 F.3d 152, 173 (2d Cir. 2001).

[9] Shyamkrishna Balganesh, The Questionable Origins of the Copyright Infringement Analysis, 68 Stan. L. Rev. 791, 805 (Forthcoming 2016).

[10] Atari, Inc. v. N. Am. Phillips Consumer Electric Corp., 672 F.2d 607, 614-15 (7th Cir. 1982).

[11] Melville B. Nimmer & David Nimmer, Nimmer on Copyright § 13.03[a][i] (2002).

[12] Scène à faire is the notion that certain similarities in the basic idea will require similarities in the expressions used to develop that idea. For example, “if two scenarios wish to treat the unprotected idea of police life in the South Bronx, one court has determined it would only be natural to depict ‘drunks, prostitutes, vermin and derelict cars,’ juxtaposed against hard drinking Irish cops chasing fleeing criminals.” Id. (citations omitted).

[13] Mark A. Lemley, Our Bizarre System for Proving Copyright Infringement, 57 J. Copyright Soc’y U.S.A. 719, 738 (2009) (alteration in original).

[14]See infra Part II.A (discussing how this need for, and lack of, special expertise in copyright makes the “ordinary observer” test a poor extension of the rationale underlying the negligence standard used in areas such as tort law).

[15] Balganesh, supra note 9, at 805 (explaining that the misappropriation inquiry requires procedural, substantive, and theoretical considerations).

[16] See generally Arewa, supra note 5 (noting that American copyright law is founded on the unrealistic conception that creativity necessarily comes from a place of pure autonomous genius).

[17] Lemley, supra note 13, at 739 (citing Ann Bartow, Copyrights and Creative Copying, 1 U. Ottawa L. & Tech. J. 77 (2003) (suggesting that this has been the result)).

[18] See generally Jeffrey Cadwell, Expert Testimony, Scenes A Faire, and Tonal Music: A (Not So) New Test for Copyright Infringement, 46 Santa Clara L. Rev. 137 (2005) (arguing functional constraints make music prone to tendencies and commonalities).

[19] See generally Alice Kim, Expert Testimony and Substantial Similarity: Facing the Music in (Music) Copyright Infringement Cases, 19 Colum.-VLA J.L. & Arts 109, 124–125 (Fall 1994/Winter 1995) (arguing that music operates by such complexities and intricacies, especially in today’s technological world as “all pieces of music, contain elements of ‘melody, harmony, . . . rhythm [,] . . . [t]imbre (tonal quality), tone, pitch, tempo, spatial organization, consonance, dissonance, phrasing, accents, note choice, combinations, interplay of instruments, . . . bass lines, and the new technological sounds.’”) (quoting Debra Presti Brent, The Successful Music Copyright Infringement Suit: The Impossible Dream, 7 U. Miami Ent. & Sports L. Rev. 229, 248–49 (1990)).

[20] See id. at 124.

[21] Callie L. Pioli, Copyright: Infringement v. Homage, (September 17, 2015, 8:11 AM), http://www.mbbp.com/news/blurred-lines-copyright (noting that the lesson from the Blurred Lines decision is that “creating music ‘reminiscent’ of an era or paying homage to the genre-creating greats of past decades may not hold as a defense to copyright infringement under the current substantial similarity framework.”).

[22] See Cadwell, supra note 18, at 127 (arguing that an analogy between music and software does not seem inappropriate and thus proposing that the substantial similarity test does not suit the technical nature of music as well as permitting expert testimony would).

[23] See generally Arewa, supra note 5, at 581, 584-85 (arguing that in viewing classical composers as artistic geniuses, failing to appreciate their practice of borrowing from the past, and hip-hop artists as mere craftsmen, legal discourse is perpetuating culturally rooted prejudices against the “other,” as most modern genres originated in African cultures).

[24] Id. at 587(“this vision of musical authorship based upon notions of creativity, invention, originality and even genius is far too restrictive a representation of musical creation.”).

[25] Harold Lloyd Corp. v. Witwer, 65 F.2d 1, 18 (9th Cir. 1933), cert. denied, 296 U.S. 669 (1933).

[26] Eriq Gardner, ‘Blurred Lines’ Jury Orders Robin Thicke and Pharrell Williams to Pay $7.4 Million, Hollywood Rep. (March 10, 2015, 2:33 PM), http://www.hollywoodreporter.com/thr-esq/blurred-lines-jury-orders-robin-779445.2.

[27] This test is currently reserved for computer software cases. See Computer Assocs. Int’l, Inc. v. Altai, Inc., 982 F.2d 693, 707 (2d Cir. 1992) (“Professor Nimmer suggests, and we endorse, a ‘successive filtering method’ for separating protectable expression from non-protectable material.”).

[28] Cf. id.

[29] Lemley, supra note 13, at 741.

[30] Balganesh, supra note 9, at 855-58.

[31] Arnstein v. Porter, 154 F.2d 464, 478 (2d Cir. 1946) (Clark, J., dissenting) (“Further, my brothers reject as ‘utterly immaterial’ the help of musical experts as to the music itself (as distinguished from what lay auditors may think of it, where, for my part, I should think their competence least), contrary to what I had supposed was universal practice . . . .”).

[32] See generally Robert Kirk Walker & Ben Depoorter, Unavoidable Aesthetic Judgments in Copyright Law: A Community of Practice Standard, 109 Nw. U.L. Rev. 343, 376 (2015) (proposing that experts brief the court on the aesthetic norms and traditions that inform the works at issue so that the hypothetical viewer is not limited to any specific aesthetic theory and can react sensitively to the nature of the work presented).

[33] See Barton Beebe, Intellectual Property Law and the Problem of Aesthetic Progress, Inaugural Lecture of the John M. Desmarais Professorship of Intellectual Property Law, NYU Law News (Feb. 3, 2014), http://www.law.nyu.edu/news/barton-beebe-inaugural-desmarais-lecture.html (discussing the courts’ failure to recognize that the framers intended to quarantine the fine arts from copyright law, as evident in the progress clause specifying “useful arts,” and the resulting lack of a developed idea of what aesthetic expression means); see also Walker & Depoorter, supra note 32, at 344-45 (explaining how courts shy away from judging art for fear that they are “incompetent to do so”) (citing Bleistein v. Donaldson Lithographing Co., 188 U.S. 239, 251 (1903)(“It would be a dangerous undertaking for persons trained only to the law to constitute themselves final judges of the worth of pictorial illustrations, outside of the narrowest and most obvious limits.”)).

[34] Michael W. Carroll, The Struggle for Music Copyright, 57 Fla. L. Rev 907, 934 (2005).

[35] U.S. Const. art. I, § 8, cl. 8 (alteration in original).

[36] See Act of May 31, 1790, ch. 15, 1 Stat. 124.

[37] William F. Patry, Copyright Law and Practice, The Bureau of National Affairs (2000), available at http://digital-law-online.info/patry/patry5.html.

[38] 8 Anne, ch. 19 (1710) (Eng.), http://www.copyrighthistory.com/anne.html.

[39] The first copyright action for a musical work, Bach v. Longman, 98 Eng. Rep. 1274, 1275 (K.B. 1777), was brought under the Statute of Anne by Johann Christian Bach on account of unauthorized editions, published by music publishers Longman & Lukey, of two Bach works, a lesson and a sonata. “Bach brought the suits seeking to effect legal changes to provide composers with copyright protection equal to that of authors.” Arewa, supra note 5, at 557–58.

[40] Arewa, supra note 5, at 568.

[41] Id. at 556.

[42] See 17 U.S.C. § 102(a)(7) (2000) (granting copyright protection to sound recordings).

[43] Paul Théberge, Technology, Creative Practice and Copyright, in Music and Copyright at 140 (Lee Marshall and Simon Frith eds., 2nd ed. 2004).

[44] Arewa, supra note 5, at 556.

[45] Id. at 625.

[46] Id.

[47] Id. at 556-57.

[48] Id. at 557.

[49] Id. at 626 n.445 (citing Aaron Keyt, Comment, An Improved Framework for Music Plagiarism Litigation, 76 Cal. L. Rev. 421, 432 (1988)).

[50] Id. at 614.

[51] See, e.g., LIVELOVEASAPVEVO, A$AP ROCKY – Wild For The Night (Explicit) ft. Skrillex, Birdy Nam Nam, YouTube (Feb. 10, 2016), https://www.youtube.com/watch?v=1eWdbMBYlH4.

[52] Youyoung Lee, OVER IT: It’s Time To Retire The Word “Haters,” The Huffington Post: The Blog (June 25, 2013, 4:50 PM), http://www.huffingtonpost.com/youyoung-lee/over-it-its-time-to-retire-haters_b_3492129.html (citing as an example the song “Players Gon’ Play,” by the all-girl group 3LW).

[53] Michael Epstein, Taylor Swift Is Being Sued for $42 Million for Singing “Haters Gonna Hate” in “Shake it Off,” Flavorwire (Nov. 2, 2015), http://flavorwire.com/545712/taylor-swift-is-being-sued-for-42-million-for-singing-haters-gonna-hate-in-shake-it-off.

[54] See Arewa, supra note 5, at 622.

[55] Carroll, supra note 34, at 949 (citing Charles Burney, A General History of Music (Dover 2d ed. 1957)); see also id. at 952 (arguing that by limiting protection to books, as opposed to single songs, the Statute of Anne was enacted only to protect “those who had advanced the cause of learning by producing books.”).

[56] C33/442 London Public Record Office (1772), reprinted in Ronald J. Rabin & Steven Zohn, Arne, Handel, Walsh, and Music as Intellectual Property: Two Eighteenth-Century Lawsuits, in 120 J. of the Royal Musical Ass’n 112, 140-45 (1995).

[57] Carroll, supra note 34, at 950.

[58] Id. (citing Pyle, C33/442 London Public Record Office, at 143).

[59] Peter Jaszi, Contemporary Copyright and Collective Creativity, in The Construction of Authorship: Textual Appropriation in Law and Literature 29, 40 (Martha Woodmansee & Peter Jaszi eds., 1994) (“Eighteenth-century theorists . . . minimized the element of craftsmanship . . . in favor of the element of inspiration, and they internalized the source of that inspiration. That is, the inspiration for a work came to be regarded as emanating not from outside or above, but from within the writer himself.”).

[60] Amy B. Cohen, Copyright Law and the Myth of Objectivity: The Idea-Expression Dichotomy and the Inevitability of Artistic Value Judgments, 66 Ind. L.J. 175, 203 (1990); see also Arewa, supra note 5, at 566 n.80 (Romantic ideals emphasize “original ideas rather than ‘successive elaborations of an idea or text by a series of creative workers.’”).

[61] Arewa, supra note 5, at 566.

[62] At least music composed in the twelve-tone scale.

[63] Arewa, supra note 5, at 590.

[64] As a consequence, “inspired work was made peculiarly and distinctively the product – and the property – of the writer.” Id. at 566-67 n.82.

[65] Arewa, supra note 5, at 592.

[66] Cohen, supra note 60, at 204.

[67] Id. at 201.

[68] Holmes v. Hurst, 174 U.S. 82, 86 (1899).

[69] See, e.g., White-Smith Music Publ’g Co. v. Apollo Co., 209 U.S. 1, 17 (1908) (holding that a perforated piano roll used to create the sounds of a musical composition did not infringe the copyright in the underlying musical composition because “[a] musical composition is an intellectual creation which first exists in the mind of the composer. . . It is not susceptible of being copied until it has been put in a form which others can see and read. The statute has not provided for the protection of the intellectual conception apart from the thing produced . . . .”).

[70] Holmes, 174 U.S. at 86.

[71] Cohen, supra note 60, at 206.

[72] Holmes, 174 U.S. at 86 (“[C]ertain words . . . are as little susceptible of private appropriation as air or sunlight[.]”); see also Johnson v. Donaldson, 3 F. 22, 24 (S.D.N.Y. 1880) (“A copyright secures the proprietor against the copying, by others, of the original work, but does not confer upon him a monopoly in the intellectual conception which it expresses.”).

[73] Cohen, supra note 60, at 231.

[74] Id.

[75] Id. at 204–206 (“If art was no longer viewed as the formal expression of fundamental, abstract ideas, but rather as the expression of the individual feelings of the particular artist, then the view that copyright should protect only the author’s specific way of expressing the ideas, but not those fundamental, abstract ideas themselves, had lost its philosophical basis.”).

[76] Act of Mar. 4, 1909, ch. 320, § 1(b), (d), (e), 35 Stat. 1175 (providing the copyright owner with the exclusive right to transform the protected work into different formats, including the right to dramatize a nondramatic work, to translate a literary work or “to make any other version thereof,” to perform works publicly, and to "make any arrangement or setting of it or of the melody of it in any system of notation or any form of record in which the thought of an author may be recorded and from which it may be read or reproduced.”).

[77] Cohen, supra note 60, at 206 (citing Benedetto Croce’s view that “the essence of artistic activity is not the production of an external physical object, but an internalized aesthetic synthesis of impressions and sensations.”).

[78] Kalem Co. v. Harper Bros., 222 U.S. 55 (1911).

[79] Id. at 63.

[80] Nichols v. Universal Pictures Corp., 45 F.2d 119, 121 (2d Cir. 1930).

[81] Nimmer, supra note 11, at § 13.03[a][i].

[82] See Cohen, supra note 60, at 207.

[83] Id. (citing John Dewey, Art as Experience 8, 49-52, 56, 64-65 (1934)).

[84] Sol LeWitt, Paragraphs on Conceptual Art, Artforum (June 1967), http://radicalart.info/concept/LeWitt/paragraphs.html.

[85] See id.

[86] Take, for example, Andy Warhol’s Brillo Boxes. By reframing the household-cleaning product as art, Warhol instilled in the Brillo box an entirely different meaning. Instead of representing a product or brand identity, Warhol’s Brillo Boxes stood for the Pop Art movement’s challenge to the dominant view of elitist aesthetics, and represented the idea that anything can be art. As philosopher Arthur Danto put it, the Brillo Boxes were the “end of art” as we know it because they marked the point at which art became so conscious of itself that it became apparent that in art “anything goes…that there were no stylistic or philosophical constraints.” Id.

[87] See Cohen, supra note 60, at 232.

[88] Id. at 212; see, e.g., Nichols v. Universal Pictures Corp., 45 F.2d 119 (2d Cir. 1930), cert. denied, 282 U.S. 902 (1931) (finding stories of star-crossed lovers too common to be protectable, despite both stories involving a relationship between a Jewish family and an Irish family, a secret marriage between the son and daughter of these two families, a conflict between the two fathers, and an ultimate reconciliation); Steinberg v. Columbia Pictures, 663 F. Supp. 706, 708–09 (S.D.N.Y. 1987) (discussing the plaintiff’s fame and the popularity of his work in finding the defendant’s work infringing); Cohen supra note 60, at 229 (“[A]nother factor that affects a court’s determination of where to draw the line between idea and expression in a given case involving literary works is the relative commercial success of the works at issue and the reputations of their creators.”).

[89] See, e.g., the recent Blurred Lines verdict discussed infraat Part II.C.1.

[90] Herbert Rosenthal Jewelry Corp. v. Kalpakian, 446 F.2d 738, 742 (2d Cir. 1982).

[91] Cohen, supra note 60, at 231.

[92] Cadwell, supra note 18, at 157; see also Arewa, supra note 5, at 556.

[93] Bruce Benward & Marilyn Saker, Music: In Theory and Practice (7th ed., McGraw Hill 2003).

[94] See generally Cadwell, supra note 18, at 155–57.

[95] Id. See generally Carol L. Krumhansl & Lola L. Cuddy, A Theory of Tonal Hierarchies in Music, in Music Perception 51 (M.R. Jones et al. eds., 2003).

[96] Cadwell, supra note 18, at 158.

[97] Id.

[98] See Yen, supra note 2, at 284.

[99] “Striking similarity” is similarity that is “so striking that the possibilities of independent creation, coincidence and prior common source are, as a practical matter, precluded.” Selle v. Gibb, 741 F.2d 896, 897 (7th Cir. 1984).

[100] Cohen, supra note 6, at 724.

[101] Id. at 728.

[102] Id. at 724–27 (comparing the use of phrases such as “substantial identity” or “substantial copy” by the courts to “signify a degree or type of similarity that would be relevant” to proving whether the defendant had in fact used the plaintiff’s work, versus the use of the adjective “substantial” in relation to the economic or aesthetic value of the copyright owner’s work to determine whether the defendant could be liable for copyright infringement.). Cf. Arnstein v. Porter, 154 F.2d 464, 473 (2d Cir. 1946) (inquiring “whether defendant took from plaintiff’s works so much of what is pleasing to the ears of lay listeners, who comprise the audience for whom such popular music is composed, that defendant wrongfully appropriated something which belongs to the plaintiff.”).

[103] Cohen, supra note 6, at 725–26 (“The focus was not principally on how much or what aspects of the plaintiff’s work defendant had borrowed, but on whether defendant had copied the plaintiff’s work rather than doing his own work. The concern was with whether ‘the labors of the original author are substantially to an injurious extent appropriated by another.’”) (quoting Greene v. Bishop, 10 F. Cas. 1128, 1134 (C.C.D. Mass. 1858) (No. 5763)).

[104] Emerson v. Davies, 10 F. Cas. 615, 622 (C.C.D. Mass. 1845) (No. 4436) (noting, in comparing the similarities between tables in two arithmetic textbooks, that “[t]he question is not in what part of one or more pages the matter is found, but whether it is borrowed or pirated from the plaintiff, without any substantial alteration or difference.”). The court went on to state that “[a] copy is one thing, an imitation or resemblance another…. It is very clear that any use of materials…which are well known and in common use, is not the subject of a copy-right, unless there be some new arrangement thereof.” Id.

[105] See Story v. Holocombe, 23 F. Cas. 171, 173 (C.C.D. Ohio 1847) (No. 13497) (stating that infringement “does not depend so much upon the length of the extracts as upon their value.”).

[106] Cohen, supra note 6, at 727 (citing Folsom v. Marsh, 9 F. Cas. 342, 348 (C.D.D. Mass. 1841) (No. 4901) (stating that infringement may exist “if so much is taken, that … the labors of the original author are substantially to an injurious extent appropriated by another”) (emphasis added).

[107] Arnstein v. Porter, 154 F.2d 464 (2d Cir. 1946). See also Nimmer, supra note 11, at § 13.03[a][i].

[108] Arnstein, 154 F.2d at 467. See also Cadwell,supra note 18, at 139 n.19.

[109] Cadwell,supra note 18, at 139 n.19.

[110] Eric C. Osterberg & Robert C. Osterberg, Substantial Similarity in Copyright Law § 3.2.1.A (PLI 2015).

[111] Arnstein, 154 F.2d at 468. See also Osterberg, supra note 110, at § 3.1.1; Cohen, supra note 6, at 729 (“Although some dispute still exists as to whether the plaintiff must prove actual access or only opportunity for access, courts generally agree that showing some possibility of access is very much a part of the plaintiff’s case.”).

[112] Arnstein, 154 F.2d at 468.

[113] Nimmer, supra note 11, at § 13.03[a][i].

[114] A piece-by-piece examination of the works’ constituent parts or elements. See Osterberg, supra note 110, at § 3.4.

[115] Cohen, supra note 6, at 731.

[116] Nimmer, supra note 11, at § 13.03[a][i].

[117] Id.

[118] See Castle Rock Entm’t, Inc. v. Carol Pub. Group, Inc., 150 F.3d 132, 137 (2d Cir. 1998). See also Osterberg, supra note 110, at §§ 3-4.

[119] Osterberg, supra note 110, at § 3.1.2 (citing Peter Pan Fabrics, Inc. v. Martin Weiner Corp., 274 F.2d 487, 489 (2d Cir. 1960)).

[120] Cohen, supra note 6, at 732.

[121] Id.

[122] Carol Barnhart, Inc. v. Econ. Cover Corp., 773 F.2d 411, 422 (2d Cir. 1985) (Newman, J. dissenting) (emphasis added).

[123] Arnstein v. Porter, 154 F.2d 464, 473 (2d Cir. 1946).

[124] See Yen, supra note 2, at 291 (“[D]ifficulties arise because the ordinary observer is not a real person whose views may be discovered.”).

[125] Arnstein, 154 F.2d at 473.

[126] Id. at 468.

[127] Nimmer, supra note 11, at §13.03[a][i].

[128] Cohen, supra note 6, at 732.

[129] Nimmer, supra note 11, at § 13.03[a][i]. The terms “comprehensive non-literal similarity” and its counterpart, “fragmented literal similarity,” emerge from this treatise but they have gained widespread judicial acceptance.

[130] Id. See also Arnstein, 154 F.2d at 473 (deeming it “an issue of fact which a jury is peculiarly fitted to determine.”).

[131] Nimmer, supra note 11, at § 13.03[a][i].

[132] Eric Osterberg, Copyright Litigation: Analyzing Substantial Similarity, 1, 7 Practical Law Intellectual Property & Technology, available at http://www.osterbergllc.com/wp-content/uploads/2013/09/Practical-Law-Article.pdf.

[133] Id.

[134] Sheldon v. Metro-Goldwyn Pictures Corp., 81 F.2d 49, 56 (2d Cir. 1936).

[135] Osterberg, supra note 132, at 7.

[136] Nimmer, supra note 11, at §13.03[a][i].

[137] Arnstein v. Porter, 154 F.2d 464, 468 (2d Cir. 1946).

[138] See Ideal Toy Corp. v. Fab-Lu Ltd., 266 F. Supp. 755 (S.D.N.Y. 1965), aff’d, 360 F.2d 1021 (2d Cir. 1966).

[139] Id. at 1022 (stating that the plaintiff need only show substantial similarity between the two works, which is present when “an average lay observer would recognize the alleged copy as having been appropriated from the copyrighted work.”).

[140] See Cohen, supra note 6, at 733.

[141] Id. at 737.

[142] The idea-expression dichotomy is essential to copyright law, as only the expression of ideas is protectable. 17 U.S.C. § 102(b) (1982) (excluding from the subject matter of copyright “any idea, procedure, process, system, method of operation, concept, principle, or discovery.”).

[143] Cohen, supra note 6, at 735 (deeming the court’s test in Ideal Toy the “Traditional Approach”) (citing Novelty Textile Mills v. Joan Fabrics Corp., 558 F.2d 1090, 1093 (2d Cir. 1977). Novelty Textile Mills found infringement because to lay eyes, the fabrics were “almost identical;” however, the court never analyzed the similarities to determine the likelihood of independent creation or the likelihood of copying.

[144] Universal Athletic Sales Co. v. Salkeld, 511 F.2d 904, 907 (3d Cir. 1975), cert. denied, 423 U.S. 863 (1975) (“[S]ubstantial similarity to show that the original work has been copied is not the same as substantial similarity to prove infringement.”).

[145] Cohen, supra note 6, at 747 (citing Salkeld, 511 F.2d at 907).

[146] Id.

[147] Id.

[148] Id. at 737 (“By relying upon the ordinary observer test alone and thus rejecting dissection, analysis, and expert testimony, the courts were deprived of the evidence necessary to analyze properly the likelihood of independent creation.”).

[149] Nimmer, supra note 11, at § 13.03[A][1].

[150] Cohen, supra note 6, at 749; cf. Nimmer, supra note 11, at § 13.03[A][1][a] (“Obviously, no principle can be stated as to when an imitator has gone beyond the ‘idea,’ and has borrowed its ‘expression.’ Decisions must therefore inevitably be ad hoc.”) (citing Peter Pan Fabrics, Inc. v. Martin Weiner Corp., 274 F.2d 487, 489 (2d Cir. 1960)).

[151] See generally Sid & Mart Krofft Television Prods., Inc. v. McDonald’s Corp., 562 F.2d 1157 (9th Cir. 1977).

[152] Cadwell, supra note 18, at 150.

[153] See Krofft, 562 F.2d at 1164 (“We believe that the court in Arnstein was alluding to the idea-expression dichotomy which we make explicit today.”); see also Cohen, supra note 6, at 753 (“The Ninth Circuit recognized that the ordinary observer is unlikely to be able to separate idea from expression in comparing two works without dissection or analysis.”).

[154] Krofft, 562 F.2d at 1164.

[155] Id. at 1165 (establishing that the question is whether the defendant took “so much of what is pleasing to the audience” to be held liable).

[156] Osterberg, supra note 110, at § 3-3-3.

[157] Shaw v. Lindheim, 919 F.2d 1353, 1360 (9th Cir. 1990).

[158] Cavalier v. Random House, Inc., 297 F.3d 815, 822 (9th Cir. 2002).

[159] Id.

[160] Shaw, 919 F.2d at 1359.

[161] Cadwell, supra note 18, at 151.

[162] Shaw, 919 F.2d at 1359.

[163] McCulloch v. Albert E. Price, Inc., 823 F.2d 316, 319 (9th Cir. 1987) (noting that a conclusion that two works are “‘confusingly similar in appearance’ is tantamount to finding substantial similarities in the objective details of the [works].”) (citing Litchfield v. Spielberg, 736 F.2d 1352, 1356 (9th Cir. 1984)).

[164] Cavalier, 297 F.3d at 826; see also Smith v. Jackson, 84 F.3d 1213, 1218 (9th Cir. 1996) (applying Shaw‘s rule to musical motifs).

[165] Cavalier, 297 F.3d at 826.

[166] Sid & Mart Krofft Television Prods., Inc. v. McDonald’s Corp., 562 F.2d 1157, 1164 (9th Cir. 1977).

[167] Arnstein v. Porter, 154 F.2d 464, 473 (2d Cir. 1946).

[168] Krofft, 562 F.2d at 1166–67. (“We do not believe that the ordinary reasonable person, let alone a child, viewing these works will even notice that Pufnstuf is wearing a cummerbund while Mayor McCheese is wearing a diplomat’s sash.”) (emphasis added).

[169] Id.

[170] See Julie E. Cohen et al., Copyright in a Global Information Economy, Companion Website, http://coolcopyright.com/contents/chapter-5/sid-marty-krofft-television-productions-v-mcdonalds.

[171] 919 F.2d 1353 (9th Cir. 1990); see also Cavalier v. Random House, Inc., 297 F.3d 815, 822 (9th Cir. 2002).

[172] Shaw, 919 F.2d at 1357 (explaining that they were comparing “every element that may be considered concrete”).

[173] Litchfield v. Spielberg, 736 F.2d 1352, 1356 (9th Cir. 1984).

[174] See Shaw, 919 F.2d at 1357.

[175] Id.

[176] Id.

[177] Swirsky v. Carey, 376 F.3d 841, 844–45 (9th Cir. 2004).

[178] Id. at 843.

[179] Id. at 847.

[180] Id. at 848.

[181] Id. at 849 n.15 (quoting Metcalf v. Bochco, 294 F.3d 1069, 1073 (9th Cir. 2002) (quoting Kouf v. Walt Disney Pictures & Television, 16 F.3d 1042, 1045 (9th Cir. 1994))).

[182] Id. at 848.

[183] Three Boys Music Corp. v. Bolton, 212 F.3d 477 (9th Cir. 2000).

[184] Id. at 485.

[185] Swirsky, 376 F.3d at 850; see also Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 349 (1991) (extending copyright protection to the original selection and arrangement of otherwise uncopyrightable components).

[186] Apple Computer v. Microsoft Corp., 35 F.3d 1436, 1442, 1447 (9th Cir. 1994).

[187] Dawson v. Hinshaw Music Inc., 905 F.2d 731, 733 (4th Cir. 1990) (describing Arnstein as “the source of modern theory”); see also Lemley, supra note 13, at 719.

[188] Arnstein v. Porter, 154 F.2d 464, 468 (2d Cir. 1946).

[189] As Professor Nimmer suggests, conversely, it may also cause very real appropriation to go undetected. Nimmer, supra note 11, at § 13.03[E][2].

[190] Id.

[191] Id.

[192] Id.

[193] Id.

[194] Id.

[195] See Balganesh, supra note 9, at 794 (“Copyright’s infringement analysis has been variously described as ‘bizarre,’ ‘mak[ing] no sense,’ ‘viscid,’ and ‘problematic.’”); see also Whelan Assocs., Inc. v. Jaslow Dental Lab., Inc., 797 F.2d 1222, 1232 (3d Cir. 1986) (noting the doubtful value of the ordinary observer test in cases involving complex subject matter unfamiliar to most members of the public).

[196] Especially when applied to complex works, such as computer software and music and with complicating circumstances of the transformation of a work into a different medium. Nimmer, supra note 11 at § 13.03[E][2].

[197] Id.; cf. Hein v. Harris, 175 F. 875, 876 (C.C.S.D.N.Y. 1910) (explaining that the ultimate inquiry is whether the average person’s ear would find the two melodies substantial similar because the pecuniary value of a composition rests in the public taste).

[198] Michael Der Manuelian, The Role of the Expert in Music Copyright Infringement Cases, 57 Fordham L. Rev. 126, 131 (1988) (describing how the trier must determine, not his own