The Court of Appeals for the Federal Circuit is the nations'preeminent patent law court. But curiously, it sometimes decides important copyright law cases too. The primary way that copyright cases reach the Federal Circuit is when a copyright case from another circuit also includes a patent claim, since all appeals of patent claims go to the Federal Circuit. Yet by the time many of these copyright-patent cases reach the Federal Circuit, the patent claims are dead, and the appeal only concerns copyright law issues. Commentators have worried about the Federal Circuits'involvement in important copyright cases because of a perception that the court is biased in favor of rights holders, lacks copyright law expertise, and may create significant legal uncertainty with its copyright law decisions. These factors may also encourage copyright plaintiffs to forum shop their copyright appeals to the Federal Circuit by including a trivial patent claim in their case. Scholars have responded to these and related concerns by calling for greater scrutiny of the Federal Circuits'non-patent case law. Despite these calls for action, we only have anecdotal accounts of the Federal Circuits'involvement in copyright law. In this Article, we take up the task of assessing the Federal Circuits'role in copyright law. As part of that assessment, we review all of the Federal Circuits'available copyright law opinions to learn more about the Court as a copyright law decision-maker. That review helps inform our analysis of the courts' relative advantages and disadvantages as a copyright law court. Overall, we conclude that despite concerns about the courts'involvement in important copyright law cases, the Federal Circuit is relatively well equipped to handle them. While the Federal Circuit may sometimes engage in copyright law mischief, its relative advantages outweigh its disadvantages and provide it with a stable foundation for a productive role in copyright law and policy going forward. However, the Federal Circuits'growing involvement with software copyright cases may change the calculus significantly if the court becomes the de facto "supreme court" of software copyright law appeals, because such a role may make the court more disposed to formalistic, error-prone decision-making in that sphere. That outcome, in turn, would undermine copyrights'constitutional purpose of advancing societal progress.

Creative Commons (CC) licenses are a series of open licenses that provide a simplified method for creators (i.e., copyright holders) to license materials in a way that is more open to the public. The three-layer design simplifies the CC licenses while still providing versions that can be read and used by lawyers and computers. The six CC licenses are: CC BY, CC BY-SA, CC BY-ND, CC BY-NC, CC BY-NC-SA, and CC BY-NC-ND. Other tools are the CC0 and Public Domain Mark.

The Digital Millennium Copyright Act (“DMCA”) has been a flash-point during most of its twenty-five-year existence. One of the most controversial parts of the DMCA is section 1201. Among other things, section 1201 prohibits third parties from circumventing certain con-trols to copyrighted content or trafficking in tools that enable circum-vention of technological controls. However, despite its nearly quarter-of-a-century lifespan, we know very little about section 1201 empiri-cally. While empiricists have assessed parts of the DMCA, they have left section 1201 largely untouched. Our understanding of section 1201 is largely based on anecdotal evidence, in the form of leading opinions from historically prominent copyright circuits. In this Arti-cle, we seek to bolster these anecdotes with greater empirical evidence for ongoing discussions about how section 1201 is performing and whether it needs revising. To do so, we conducted a broad-based search of Westlaw to collect every issued opinion, whether reported or not, where a court purported to apply some part of section 1201. We then reviewed these cases to glean as much information about section 1201 as possible. This re-view led to a number of important and, in some cases, surprising re-sults. First, section 1201 opinions are a relative rarity. In the nearly quarter of a century since the DMCA’s enactment, we could find only a little over 200 opinions, with only about sixty of those being published. The average number of opinions during the DMCA’s existence has been around nine annually, which pales in comparison to other types of copyright cases. Second, despite the Second Circuit receiving much attention in anecdotal accounts of section 1201, courts within it issue section 1201 opinions infrequently. The Ninth Circuit is the dominant section 1201 court, both in terms of citations to its opinions and over-all number of opinions, and the Sixth and Eleventh Circuits both is-sue more section 1201 opinions than the Second Circuit. This result stands in contrast to other types of copyright litigation, where the Sec-ond Circuit is a behemoth. Third, the most common subject matter in dispute in section 1201 cases is computer software, followed distantly by audiovisual material such as movies. Music stands in last place, showing up in only a couple issued opinions. Debates at the time of the DMCA’s enactment were informed by widespread fears of copy-right infringement relating to digital music and other types of digital content. Yet section 1201 litigation has resulted in but few written opinions involving those subject matters. Fourth, suits and defaults against individuals happen relatively frequently in the section 1201 context, with courts often assessing large statutory damages against those individuals. As we discuss in this Article, this result raises im-portant equity issues. Fifth, despite section 1201 including a number of statutory exemptions, these exemptions basically never make their way into issued opinions. Fair use, too, only infrequently enters courts’ section 1201 discussions. This means, effectively, that the primary way to escape section 1201 liability is through administrative exemp-tions granted by the Library of Congress on a triennial basis. But as we shall see, this process has significant holes. Finally, plaintiffs dis-proportionately win section 1201 cases. This result is somewhat bloat-ed because of the frequency of defaults against individuals. Setting these aside, plaintiffs still enjoy tremendous success under section 1201. However, when looking at opinions only outside of the Ninth Circuit, win rates become mostly even. I conclude with several calls for DMCA reform. These include bol-stering statutory exemptions and more closely tying section 1201 to copyright infringement. Pursuing these reforms will more faithfully align section 1201 with its purported objectives.

ion-filtration-comparison test rises significantly to nearly a quarter of cases. This almost certainly underestimates the use of the abstraction-filtration-comparison test in software cases, though, as this study defined technology to include contexts beyond simply software. Indeed, as Professor Pamela Samuelson notes, nearly every circuit has used the abstraction-filtration-comparison test for assessing claims of software copyright infringement.207 Our data largely confirm Samuelson’s findings; our study includes multiple opinions utilizing the abstraction-filtration-comparison test from each of the first eleven circuits, omitting only the Federal Circuit and D.C. Circuit, where copyright cases rarely happen. Aside from these three traditional groupings, our data highlight several additional insights regarding prong two. First, the third most popular means by which courts decide prong two is to use no specific subtest at all. Nearly 24% of all opinions purporting to reach a result on prong two did not clearly apply any of the subtests. That result might seem curious, especially in light of how many options courts have. Several factors help explain this outcome. For starters, as we shall see shortly, a good number of courts simply rely on copyright limitations to determine whether substantial similarity under prong two exists; in essence, they use these limitations as the subtest for assessing similarity. Furthermore, a number of our opinions come at relatively early stages of litigation, where the court simply assesses whether the plaintiff has pled enough to state a legal claim under the substantial similarity standard. In such scenarios, courts sometimes do not engage in a deep analysis of the substantial similarity question, omitting discussion of specific subtests and copyright limitations in making their determinations. Finally, sometimes courts are simply sloppy and conflate prongs one and two. For instance, at times they decide prong two largely on the basis of prong-one considerations, concluding that the defendant clearly copied from the plaintiff when, at least under prong two, that is not the pertinent question.208 These three factors almost certainly contribute to the near quarter of opinions where courts failed to apply a prong-two subtest when determining the prong-two outcome. A second important finding is that courts frequently use a number of other subtests when deciding prong two, not just the three headliners. For instance, the “total concept and feel” subtest is standard fare in many opinions. It is frequently used in the Second Circuit in conjunction with the ordinary observer test and in the Ninth Circuit as part of the intrinsic part of the extrinsic-intrinsic test.209 This subtest asks courts to assess similarity based on the overall aesthetic of the two 207. Pamela Samuelson, A Fresh Look at Tests for Nonliteral Copyright Infringement, 107 NW. U. L. REV. 1821, 1837–38, 1838 n.108 (2013). 208. See, e.g., Gracing Inc. v. E.K. Blue, Inc., No. CV 16-5107, 2017 WL 5640516, at *1–3 (C.D. Cal. June 28, 2017) ( seeming to conflate the prongs one and two questions by emphasizing that independent creation was highly unlikely during its discussion of prong two). 209. Hamil, America Inc. v. GFI, 193 F.3d 92, 102 (2d Cir. 1999) (assessing total concept and feel as part of the ordinary observer test ); YS Built LLC v. Ya Hsing Chiang Cind Huang, 739 F. App’x 414, 416 (9th Cir. 2018) (applying total concept and feel as part of the intrinsic test ). Second to Printer_Asay.docx (Do Not Delete) 1/5/2023 3:13 PM 84 UC IRVINE LAW REVIEW [Vol. 13:35 works rather than simply dissecting them and only focusing on the protectable pieces in isolation.210 Overall, courts assessed works’ similarities in terms of total concept and feel in a little over 21% of this study’s opinions, making it the fourth most popular subtest under prong two. Perhaps less expected is the rise of the use of some form of the “striking similarity” subtest under prong two. Courts used this subtest under prong two in about 18.5% of our opinions, making it the fifth most popular subtest under prong two. Furthermore, within our study courts from every circuit except the Federal Circuit issued opinions that relied on striking similarity to resolve prong-two questions. These outcomes are somewhat surprising because, as discussed above, traditionally courts have used this subtest under prong one to infer access when evidence of access is lacking. The idea, at least under prong one, is that the similarities between the two works are so significant that the court simply can’t fathom that one wasn’t copied from the other.211 But as we have discussed, courts mostly don’t assess similarities under prong one, at least anymore. Instead, comparison of the works, to the extent that it happens, typically occurs under prong two. This makes use and migration of the striking similarity subtest under prong two much more understandable, if not justified. Indeed, some courts that use striking similarity under prong two seem to use it as if answering prong one: the similarities are so striking that it is reasonable to believe that the defendant copied from the plaintiff.212 Yet that is not the question under prong two, where courts are supposed to assess whether what was copied was a substantial amount of protectable material. However, at least some courts seem to imply that when striking similarities exist, that is enough to satisfy prong two as well, since the striking similarities are likely to encompass whatever protectible material exists in the plaintiff’s work.213 Some circuits use a form of the striking similarity test under prong two primarily when the plaintiff’s work only enjoys “thin” copyright protection, meaning that the work consists of mostly unprotectable material. In such scenarios, courts within the Ninth Circuit insist that prong two is satisfied only if there is “virtual identity” between the two works.214 Again, the idea seems to be that such striking similarities will inevitably encompass whatever little protectable material the plaintiff’s work possesses, though courts in such scenarios typically avoid specifically identifying what in the work is actually protectable. Courts in the Tenth 210. Hamil America, Inc., 193 F.3d at 102; YS Built LLC, 739 F. App’x at 416. 211. Gracing Inc., 2017 WL 5640516, at *1–3. 212. Sweet People Apparel, Inc. v. Fame of NY, Inc., No. 11-1666, 2011 WL 2937360, at *1 (D.N.J. July 19, 2011) ( finding a likelihood of success on the merits because the “striking resemblances” between the two works defied defendants’ claims that their work had been copied from the public domain). 213. Id. 214. George S. Chen Corp. v. Cadona Int’l, Inc., No. SACV 04-365, 2006 WL 8450995, at *6 (C.D. Cal. Feb. 14, 2006). Second to Printer_Asay.docx (Do Not Delete) 1/5/2023 3:13 PM 2022 ] COPYRIGHT’S SUBSTANTIAL SIMILARITY TEST 85 Circuit similarly sometimes use what they call the “supersubstantial similarity” test, where they deem that in cases involving a copyrighted work with only thin copyright, the defendant’s work must be nearly identical to the plaintiff’s work to find infringement.215 Other courts, regardless of whether the copyright in the plaintiff’s work is thin or robust, simply note that the similarities between the two works are striking enough to satisfy prong two.216 The rise of using some form of the striking similarity subtest under prong two is not problematic on its own. It certainly makes some sense that when striking or extensive similarities exist between two works, often appropriation of protectible material will follow. Yet simply stopping the analysis at observing striking similarities, without specifically identifying protectable material that a defendant appropriated, runs the risk of finding infringement when no or little protectable material is actually in play. This risk may be particularly pronounced when courts rely on and use reasoning from other courts that employ the striking similarity subtest, since much of that case law will have used striking similarity as a means to address factual copying under prong one. A third finding is that, contrary to what others have found,217 courts frequently use multiple subtests to answer prong two. Naturally, our differences with other studies on this score reflect our decision to forego lumping the subtests together under general headings. However, for the reasons discussed above, we think this decision is justified. A little over 65% of our opinions purporting to apply subtests under prong two relied on at least two subtests when deciding that prong. In some opinions, courts relied on seven and even eight subtests under prong two. While some may argue that these percentages are inflated because in many cases courts are likely to have meant the same thing despite slight differences in terminology, the reality is that many courts used those differences in terminology somewhat consistently. That consistency suggests these differences are not always simply the result of sloppiness nor that courts always or even typically mean the same thing when using differently denominated subtests.218 Our last significant finding under prong two concerns win rates. Plaintiffs had much less success under prong two than they did under prong one. Under prong 215. Craft Smith, LLC v. EC Design, LLC, 388 F. Supp. 3d 1385, 1401 (D. Utah 2019), aff’d, 969 F.3d 1092 (10th Cir. 2020). 216. Koontz v. Jaffarian, 617 F. Supp. 1108, 1113–15 (E.D. Va. 1985). 217. Lim, supra note 8 ( finding that courts rarely use multiple subtests ). 218. Furthermore, despite trying to capture each distinct subtest in our review of opinions, in some cases we also ended up lumping slightly differently named subtests under the same group when we thought doing so was advisable. For instance, the “layperson” group includes opinions where the court referred to “layman,” “lay listeners,” “lay observers,” and oth

The software industry's history is also its future. Its history has been defined by both abundance and scarcity, and its future will be, too. In the 1970s and 80s, perceived software scarcity led U.S. legislators to formally grant intellectual property protections to software creators. Later, a different kind of scarcity—a lack of access to source code—led the founders of the free and open source software movement to flip intellectual property protections on their head in an effort to better promote abundance. That movement proved wildly successful, with today's software industry based on vast amounts of freely available open source software resources that both organizations and individuals collaboratively build. Abundance and scarcity will also define software's future, but in different ways. The abundance that the open source software movement spawned is in the midst of a significant commercial phase. That sometimes means that commercial competitors bring to the table a scarcity mindset that conflicts with the norms that made that movement so successful. Intellectual property concerns at times derail what may otherwise be even greater software abundance. And because so much software is moving into the Cloud, trade secrecy may become the software industry's most important form of intellectual property to the extent the industry abandons open models of innovation. The software industry's growing dependence on artificial intelligence (AI) is likely to contribute to these trends. The software industry is increasingly becoming synonymous with the AI industry, as more and more software companies either rely on AI in running their services or provide AI products to the public. As with all software, these AI technologies are increasingly provided from the Cloud, where trade secrecy is not only possible, but often preferable. But trade secrecy may be even more likely in the AI context because much of the magic in implementing AI systems lies in the know-how to piece them together from available open source software resources, decades-old AI techniques, and data. Hence, to the extent that software and AI technologists spurn open innovation in favor of a scarcity mindset, trade secrecy is likely to become its dominant form of legal protection. The advent of web3 technologies may eventually change some of these trends. But for now, increasing secrecy seems the most likely outcome. I conclude by arguing that this shift to secrecy is likely preferable to other forms of intellectual property.

In previous work, we explored how certain characteristics of adversity are often more conducive to innovation than others. In this Article, prepared as part of the Lee E. Teitelbaum Utah Law Review Symposium—The Law & Ethics of Medical Research, we review some of that work and apply it specifically to the COVID-19 context. We conclude by assessing certain policy implications in light of how the COVID-19 pandemic has both spurred and hindered innovation.

For nearly half a century, the United States has been one of the main proponents of harmonizing the world’s copyright laws. To that end, the U.S. government has worked diligently to persuade (and, in some cases, bully) most of the world’s countries to adopt copyright standards that resemble those found in the U.S. The primary reason for this push to harmonize the world’s copyright laws is simple: the U.S. has long been a net exporter of copyrighted works, and so the U.S. government has sought to ensure that other countries provide U.S. authors with the same economic rights those authors enjoy at home. But that rather simple calculus in favor of copyright harmonization has changed. Today, the U.S. government must also take into account the interests of its technology sector in determining its positions on both domestic and international copyright law efforts. This is because technology providers are also a significant export of the U.S., and their copyright interests do not always align with those of large copyright owners. For instance, many of those technology companies, including Google and Twitter, use copyrighted works as a vital part of providing their technological services, including by way of exhibiting copyrighted content at the direction of their users, in response to user searches, or as part of services such as Google News. Consequently, continuing to ratchet up worldwide copyright standards through international harmonization may often negatively affect the interests of such companies by restricting their ability to liberally use copyrighted works within their services. In short, with these new technological entrants, the political economy of copyright harmonization has significantly changed, and those changes are poised to exert considerable influence on copyright’s global future. In this Article, I map out the key new players in copyright’s new political economy, and I grapple with how their often divergent interests are likely to affect global copyright law and policy making going forward. I then examine the European Union’s newly minted Copyright Directive as an example of these divergent forces at play within the U.S., Europe, and elsewhere. Finally, I assess whether the altered political economy of copyright harmonization is a positive or negative development.

Adverse experiences, like long-term poverty, can inhibit innovation. But as much research and many real-world examples show, adversity can also stimulate innovation in some circumstances. Indeed, the COVID-19 pandemic provides a number of recent examples where adverse conditions have led individuals, firms, and governments to innovate in the hopes of benefiting society.Despite the fact that some forms of adversity undermine innovation while others stimulate it, legal scholars have largely failed to distinguish between the two or even account for adversity’s relationship to innovation when assessing innovation law and policy, including intellectual property (IP) laws. Yet given adversity’s significant role in affecting the pace and direction of innovation, doing so is crucial. In this Article, we undertake that task. Our analysis shows that adversity is most likely to stimulate innovation when it satisfies what we call the Goldilocks principle: the adversity is neither too intense nor too mild, too fleeting nor too enduring, too all-encompassing nor too confined, too commonly experienced nor too isolated, too severe nor too insignificant, but instead is “just right.” Hence, for adversity to have the best chance of stimulating innovation, it should be (1) a relatively discrete experience; (2) of moderate intensity; (3) experienced collectively rather than in isolation; and (4) significant enough that, left unaddressed, the adversity could result in severe consequences for large groups of people. To be clear, these conditions are not necessary for innovation—adversity, or some other trigger, might spur innovation even if each of these conditions is not met. Neither are they sufficient—innovation will not necessarily occur even if all of these conditions are present. Indeed, individual and organizational characteristics often play a role in determining whether a party will respond to adversity with innovation. But existing research suggests that these are some of the adverse conditions most conducive to, and thus most likely to inspire, innovation. Conversely, adverse conditions falling outside of these parameters are more likely to inhibit innovation, or at least fail to stimulate it.We then assess what this means for IP laws and innovation policy more generally. Predominant theories suggest that IP laws are meant to incentivize parties to benefit society through innovation and creativity. Yet over the years, commentators have pointed out that IP rights are often unnecessary to inspire these activities and thus at times impose unnecessary costs on society by restricting access to those innovations. We contribute to this important discussion by highlighting the role that adverse conditions frequently play in affecting the pace and direction of innovation. First, we argue that the role of adversity in stimulating certain innovations provides another reason to doubt the efficacy of IP rights as applied to many of those innovations. Other policy levers, such as grants and prizes, may often be preferable in such cases. Second, we explore possible solutions to innovation-inhibiting adversity, including bolstering IP rights in certain situations and a greater societal commitment to basic research funding. Finally, we examine the role that adversity can play in creating innovation path dependencies, and we briefly explore some possible solutions to this dilemma.

Copyrightability is the key issue in the SAS v WPL case, in which SAS claims that WPL infringed the copyright in its statistical analysis program because of similarities in the input and output formats of the WPL program that allows users of the SAS Language to run those programs on the WPL as well as the SAS platform. The brief asks the Federal Circuit to carefully distinguish among three different meanings of “copyrightability.” A first meaning concerns whether a creation is a “work of authorship” eligible for protection as statutory subject matter. A second meaning is whether a statutorily eligible work satisfies the constitutional requirements of original authorship and fixation. A third meaning addresses whether the specific elements of a copyrighted work alleged as the basis of infringement are within the scope of protection that copyright law provides to that work. The last meaning is at issue here: whether the nonliteral elements that SAS claims WPL copied constitute copyright-protectable expression. Longstanding precedent holds that the plaintiff’s burden to prove infringement includes a burden to prove that the individual copied elements are protectable in this third sense. Although copyright registration certificates may be prima facie evidence of a work’s copyrightability in the first two senses, they are not prima facie evidence in relation to the third sense.In keeping with cases that address copyrightability in the third sense, numerous precedents recognize that software embodies many unprotectable elements and that courts must “filter” out those elements before assessing whether infringement has occurred. This is why plaintiffs carry the burden to prove the defendant copied protectable expression in this third sense of copyrightability. The abstraction-filtration-comparison (AFC) test, which the Fifth Circuit has adopted, is essential to determining which nonliteral elements of computer programs constitute protectable expression. The abstraction and filtration steps are best understood as issues of law. The filtration step excludes unprotectable elements of the plaintiff’s work so that the comparison step focuses only on whether the defendant copied a substantial quantum of protectable expression. When copyright owners fail to identify specific protectable expression as part of this step, it is proper for courts to dismiss their claims. In this case, the district court faithfully applied the AFC test. Its copyrightability hearing was an appropriate proceeding through which to assess the AFC test’s first two steps as matters of law. Because SAS failed to engage in filtration by identifying specific protectable nonliteral elements in response to WPL’s uncopyrightability evidence, the district court properly dismissed its complaint. The AFC test also advances copyright law’s constitutional purposes and creativity-enhancing objectives. Allowing second comers to emulate existing program’s functionalities and enabling compatibility on a different platform can unleash new rounds of creativity and allow users to continue to enjoy the benefits of their creations. Copyright should not be interpreted to give SAS control over a computer language so many have learned to use. Limiting the scope of copyright protection for computer programs is essential to advancing the constitutional purpose of promoting progress in keeping with the Supreme Court’s recent pronouncements in Google v. Oracle.

Scholars have long debated whether the outputs of AI systems should be subject to copyright. On the one hand, the automated nature of many AI systems may make copyright unnecessary as an incentive for the creation of those AI systems’ outputs, in which case society would be better off withholding copyright protections from them. On the other hand, those outputs often exhibit sufficient creativity to merit copyright protection, and without copyright, parties that use AI systems to create such outputs may lack the necessary incentives to do so. In this Essay, prepared as part of the Florida International University Law Review's symposium on intelligent entertainment, I argue that copyright law’s independent creation defense, as well as the widespread availability of AI systems for helping authors in their creative efforts, help address some of the concerns embedded in these debates. Historically, the independent creation defense has rarely applied, simply because independent creation of similar expression is highly unusual. But as this Essay explores, AI increases the likelihood of multiple parties creating similar expression independently, meaning that the defense can help defuse worries that applying copyright to AI outputs will result in a copyright quagmire. Furthermore, the availability of AI systems for assisting authors in their creative efforts means that authors have tools for more readily creating unique works that avoid many of the remaining copyright landmines. Other copyright issues linger, however, and the last part of this Essay examines some of these concerns in brief. In particular, parties may wish to use specific AI outputs in their own creative efforts, and neither the independent creation defense nor the availability of AI tools for creating something unique help address this problem. Copyright law’s fair use defense may, however, and the Essay concludes by briefly examining how.

Scholars have long worried that risk aversion can have significant negative effects in the marketplace. In the intellectual property law domain, some have worried that risk-averse actors can negatively influence the development of important intellectual property law doctrines, which can ultimately hamper innovation. For instance, risk-averse actors may frequently choose to obtain licenses for rights that the relevant laws do not actually require of them. When they do so, they inadvertently increase the scope of intellectual property rights because their risk-averse activities inform courts’ development of key intellectual property law doctrines. In this Article, prepared as part of the IP Scholars Forum at Akron Law, I look at the other side of the risk coin. In particular, I argue that early-stage companies, and sometimes later-stage companies as well, are often willing to take on significant intellectual property risks in pursuit of commercial opportunities. And by providing courts with opportunities to take head-on key intellectual property questions, these risk-taking activities, in effect, may often help counterbalance whatever negative effects the behavior of risk-averse actors entails. I examine reasons why both types of entities are often willing to take on intellectual property risks. And I review a number of examples where both early and later-stage companies have heavily influenced the development of key intellectual property law doctrines by being willing to take their intellectual property disputes to court. This review, however, highlights several reasons why early-stage companies are more dependable risk-taking entities than later-stage companies. I thus conclude by briefly assessing two intellectual property-related means by which to specifically encourage early-stage companies to continue to take on intellectual property risks.

Fair use is copyright law’s most important defense to claims of copyright infringement. This defense allows courts to relax copyright law’s application when courts believe doing so will promote creativity more than harm it. As the U.S. Supreme Court has said, without the fair use defense, copyright law would often “stifle the very creativity [it] is designed to foster.” In today’s world, whether use of a copyrighted work is “transformative” has become a central question within the fair use test. The Supreme Court first endorsed the transformative use term in its 1994 Campbell v. Acuff-Rose Music, Inc. decision. Since then, lower courts have increasingly utilized the transformative use doctrine in fair use case law. In fact, in response to the transformative use doctrine’s seeming hegemony, commentators and some courts have recently called for a scaling back of the transformative use concept. So far, the Supreme Court has yet to respond. But growing divergences in transformative use approaches may eventually attract its attention. But what is the actual state of the transformative use doctrine? Some previous scholars have empirically examined the fair use defense, including the transformative use doctrine’s role in fair use case law. But few have focused specifically on empirically assessing the transformative use doctrine in much depth. This Article does so by collecting data from all district and appellate court fair use opinions between 1991, when the transformative use term first made its appearance in the case law, and 2017. These data include how frequently courts apply the doctrine, how often they deem a use transformative, and the win rates for transformative users. The data also cover which types of uses courts are most likely to find transformative, what sources courts rely on in defining and applying the doctrine, and how frequently the transformative use doctrine bleeds into and influences other parts of the fair use test. Overall, the data suggest that the transformative use doctrine is, in fact, eating the world of fair use. This Article concludes by analyzing some possible implications of the findings, including the argument that, going forward, courts should rely even more on the transformative use doctrine in their fair use opinions, not less.

The freedom to reimplement application program interfaces (APIs) in independently written software is the key issue at stake in Oracle America, Inc. v. Google Inc. In May 2016, on remand a jury found Google’s use of the Java API elements was a fair use. After the District Court declined to vacate the jury verdict, Oracle appealed for the second time to the Federal Circuit, claiming that it is entitled to a judgment in its favor on the fair use issue as a matter of law. This Article, which is a derivative work of the amicus curiae brief of 42 Intellectual Property Law Professors submitted in that case, challenges Oracle’s arguments that as a matter of law, no reasonable jury could have upheld Google’s fair use defense. The Article also highlights important considerations for future courts in the software fair use context. Although most cases testing the legality of unlicensed reimplementations of APIs have been decided on copyrightability grounds, the Oracle case tests the viability of fair use as a defense to claims of copyright infringement for API reuses. Oracle’s attack on the jury’s fair use verdict rests on numerous flawed assertions that, if accepted by the Federal Circuit, could undermine robust software innovation by severely restricting the fair use defense’s applicability in software copyright cases. How the Federal Circuit decides Oracle’s appeal of a jury verdict in favor of Google’s fair use defense will have significant implications for future software copyright fair use cases because Oracle, in effect, calls into question the viability of fair use defenses in all API reuse cases (and perhaps in software cases more generally). Fair use in the digital age has come to play an important role in balancing the interests of first- and second-generation creators in software as well as other creative fields. So, it would be not just unfortunate, but possibly devastating to competition and ongoing innovation in the software industry if fair use defenses were precluded in API reuse cases.

Parties frequently obtain patents for one purpose, only to use those patents for another. This Article calls such divergences between parties' initial motivations to obtain patents and those patents' predominant uses later on "patent schisms." Because traditional patent law theories typically treat the purposes of patents as static, scholars have neglected to explicitly examine patent schisms and the reasons behind them. This is so despite the pervasiveness of patent schisms in a variety of important contexts. Those contexts include the patenting behaviors of early-stage companies, later-stage companies, so-called "patent trolls," and universities. In fact, patent schisms lie at the heart of some of the most controversial patent law topics, including whether patents should be considered a form of personal property or, instead, as a regulatory right. This Article examines patent schisms and adds to the patent literature in three principal ways. First, it provides an account of the ubiquity of patent schisms in a variety of important settings. Second, it articulates three theories explaining how and why patent schisms arise. These hypotheses include the proposition that patenting an invention often creates economic and psychological incentives to ultimately use that patent in defiance of a party's original motivation to obtain the patent. Finally, the Article examines the normative and theoretical implications of the pervasiveness of patent schisms and the explanations behind them. These include briefly assessing whether treating patents as a form of personal property is the correct approach to ensuring that the patent system serves its constitutional purpose of promoting the progress of "science and the useful arts."

INTRODUCTIONIn 2012, Cindy Lee Garcia sued Google Inc. for infringement.1 Her lawsuit was meant to force the company to remove from YouTube an antiIslamic film that included a five-second performance by her.2 Garcia claimed a interest in the performance and that YouTube, therefore, had no right to host it without her permission.3 For $500, Garcia had agreed to the performance with the understanding that it would be used in a film called Desert Warrior.4 But when her performance was distorted and used in an antiIslamic film, Innocence of Muslims, the consequences were severe.5 When the film appeared on YouTube and elsewhere, Garcia received death threats.6 Some even suggested the attack on the U.S. embassy in Benghazi, Libya, was in response to the film.7After the district court held against Garcia, the Ninth Circuit Court of Appeals initially ruled in her favor, holding that Garcia had likely met her burden of demonstrating a interest in her performance.8 This was so because, among other reasons, her performance included some amount of creativity, even if it was only five seconds long and based on a script provided to her.9 According to the court, that creativity may include her body language, facial expression and reactions to other actors and elements of a scene.10 But in early 2015, a full panel of the Ninth Circuit reversed this earlier decision, observing that [t]reating every acting performance as an independent work [subject to copyright] would not only be a logistical and financial nightmare, it would turn [a] cast of thousands into a new mantra: of thousands.11Now fast forward to June 29, 2015. On that day, the U.S. Supreme Court denied Google's petition for writ of certiorari requesting review of the Court of Appeals for the Federal Circuit's decision in Oracle v. Google.12 As a result, the Federal Circuit decision, which upheld protection for certain parts of Oracle's Java software technologies, was left intact.13 In particular, Google's use of thirty-seven of Oracle's Java application programming interfaces (APIs) in its Android operating system may constitute infringement because, according to the decision, creation of the APIs required some creativity.14 And this was so despite the fact that the APIs were, quantitatively, only a very small part of Android; Google engineers wrote nearly all ten million lines of the software code for Android.15 Although the district court found on remand that Google's use of the APIs constituted fair use, the Federal Circuit's decision regarding the copyrightability of the APIs remained otherwise undisturbed.16While these two cases have many obvious differences, they highlight a similar potential problem: what the Ninth Circuit in Garcia called copyright of thousands. More traditionally, this type of problem is referred to as an problem, which is shorthand for underuse of a resource because numerous parties have rights in the resource, and the presence of these multiple rights inhibits others from using that resource in socially beneficial ways.17 Scholars often argue that anticommons result when any one of these rights holders asserts its rights to prevent others from using the common resource.18 But while rights assertions may be one common cause of anticommons problems, this Article takes the position that anticommons can also result when rights remain unasserted, or even, in some cases, when they are licensed. In other words, a multiplicity of rights in a resource can still result in underuse of that resource, even without formal bargaining breakdowns. This Article focuses on potential anticommons problems in the software context and argues that at least two significant and growing trends in modern software innovation are leading to rising anticommons concerns.First, because increasingly more software products are collaboratively built by a variety of parties,19 any given software product may be subject to hundreds, and sometimes even thousands, of interests. …

Oracle's attack on the jury's fair use verdict is based on at least four flawed assertions. The first is that Google's failure to license certain elements of the Java Application Program Interface (API) for its Android platform was evidence of bad faith as a matter of law. This assertion is plainly inconsistent with the Supreme Court's decision in Campbell v. Acuff-Rose Music, Inc., in which the Court not only expressed skepticism about whether subjective "faith"—good or bad—is relevant in fair use cases, but also explicitly stated that seeking, but not obtaining, a license to use another's copyrighted material is not evidence of bad faith.

Traditional patent law theories teach that a patent's rights of exclusion are a patent's key benefit to the patentee and are necessary to make the patent system work. Yet patentees are increasingly giving away such rights, in whole or in part, as part of a growing phenomenon: patent pledges. In these scenarios, patentees voluntarily commit limit enforcement of their rights. This phenomenon seems to contradict traditional patent law theories. After all, if exclusive rights are necessary, why are patentees increasingly sacrificing some or all of those rights? This chapter argues that they do so because in patent pledging contexts, patents entail a different value proposition than traditional patent law theories posit. That is, patent pledgers use patents as informational tools to signal to the relevant public their development preferences and activities. This information may then facilitate a variety of economic motives behind such pledges. This chapter reviews several patent law features that make patents valuable as informational tools, as well as others that limit their informational potencies. It concludes by assessing some implications of this informational account of patents, in particular with respect to "open innovation."

Fair use is copyright law’s most important defense against claims of copyright infringement. It provides courts with an equitable tool for allowing parties to use the copyrighted materials of others without liability when doing so facilitates copyright’s constitutional purpose of promoting the “progress of Science and the useful Arts.” When analyzing fair use, modern courts place great emphasis on whether the purportedly fair use involves a “transformative use” of the copyrighted materials. In what some are calling the most important software copyright case in decades, a jury recently handed Google a victory by concluding that Google’s reuse of some of Oracle’s Java software in its Android platform constituted fair use. Oracle has appealed the decision, claiming, among other things, that Google’s use of its copyrighted Java software could not be fair use because the use was not transformative. The use was not transformative, claim Oracle and its amici, because Google uses Oracle’s software for the same functional purposes for which it was written. This Essay contends that accepting Oracle’s argument would mean that fair use rarely if ever applies in the software reuse context. Software’s functional nature, after all, means that reuses of software will inevitably involve the software carrying out the same functions for which it was designed. But if that reality alone forecloses the possibility of a fair use defense, it means the productive balance that fair use helps strike between copyright holders and follow-on software innovators may be imperiled. The Essay concludes with some thoughts on how best to apply the transformative use inquiry in software reuse cases.

Over the last decade, much of the patent law literature has focused on the problem of "patent trolls," or patent owners who don't make products, but sue others that do. The basic complaint against these types of entities is that they impose a tax on innovation without providing offsetting societal benefits. Furthermore, their patent assertions have been on the rise, with a significant percentage of patent suits now attributable to them. In short, the troll phenomenon suggests a problem of excessive patent assertions.But despite the importance of the troll phenomenon, the fact remains that most patents are never asserted, or are asserted less than they could be. Under-assertion of patents thus appears to be more prevalent than over-assertion. Yet, beyond noting a set of generic economic considerations that may lead to this outcome, the literature fails to provide systematic, industry-specific assessments of why patent owners choose to forego asserting their rights in so many cases. And the generic nature of these assessments is particularly problematic given that patents play significantly different roles from one industry to the next, as scholars have noted for some time.This Article addresses these issues by providing an industry-specific, informal model for theorizing why patent owners forego asserting their rights in so many cases (and why they may not in others). It briefly applies this model to four industries: software, pharmaceuticals, biotechnology, and semiconductors. The Article then explores some potential implications of this industry specific model. In particular, this Article suggests that high barriers to patent assertion in an industry may, ironically, result in increased patent trolling in the industry. Hence, this Article provides guidance to policymakers by helping explain the rise of patent assertions in some industries, such as software, as well as helping to identify other industries, such as biotechnology, that may be increasingly at risk of patent trolling.