The Supreme Court’s Decision in Google v. Oracle: Transformative Use of Popular Code Can Be ‘Fair Use’

April 9, 2021

The battle over Google’s use in its Android platform of a portion of Oracle’s Java SE software platform was more than a decade in the making before the Supreme Court put it to rest earlier this week.

Oracle claimed that Google’s conduct infringed its copyright and entitled it to billions of dollars in compensation.  For its part, Google argued that what it used was not copyrightable in the first place, and that in any event its acts qualified as fair use and so were shielded from liability.  Google won on both arguments in the district court but lost in the Federal Circuit, which ruled that the software at issue is copyrightable and Google’s use of it was not fair.

After granting review of both issues, the Supreme Court ruled for Google on its fair use defense in Google LLC v. Oracle America, Inc.[1]  In doing so, the Court passed on the opportunity to issue its first-ever definitive ruling delineating when software is or is not copyrightable.  Instead, in a 6–2 decision authored by Justice Breyer, the Court assumed for purposes of argument that the code at issue—roughly 11,500 lines of “declaring code” from the Java SE Application Programming Interface—was copyrightable.  But it ruled that Google’s use of the code to enable Java-trained programmers to easily write programs for the Android mobile platform qualified as fair use—based principally on its finding that using the code for a mobile platform was transformative and that the code’s value derived more from the coding community having become familiar with it than from the type of expressive features that copyright is intended to protect.  In so ruling, the Court underscored that the fair use doctrine “can play an important role in determining the lawful scope of a computer program copyright” and that, while jury findings of fact may be given weight, the doctrine’s application is ultimately a legal question to be decided by the trial judge and reviewed de novo on appeal.

Background

The doctrine of fair use that served as the foundation for the Supreme Court’s ruling generally permits limited use of copyrighted material without having to acquire permission from the copyright owner.  Originally judge-made and then subsequently codified in the Copyright Act,[2] the doctrine aims to balance the interests of owners of copyrighted material against the public interest in the wider distribution and use of creative works by allowing certain uses that may otherwise be considered infringing.  It has been applied, for example, as a defense to infringement claims brought against parodies of copyrighted works.[3]  In recent years, fair use has been applied broadly to a variety of uses, including computer-based analytical processes.[4]  

The Java SE program at issue in the suit is a popular open-source platform for programmers to write programs in Java, one of the most commonly-used software languages.  A feature of the Java SE platform (and other programming platforms) is something called an “application program interface” (or “API”), which is pre-written code for performing various standard functions that can be invoked by programmers creating applications.[5]  An example discussed in the Court’s opinion would be code that performs the function of taking two integers and determining which is greater.  Such pre-written code saves the programmer substantial time and effort when she designs programs.  In the Java SE platform, this pre-written code is referred to as the Sun Java API (named after the original developer of Java SE, Sun Microsystems, Inc. (“Sun”)).[6]  The API, in turn, involves three categories of code: “implementing code,” which executes specific prewritten tasks; “method calls,” which select the set of implementing code for a certain task; and “declaring code,” which specifies the name of each task and the location of each task in the API, thereby allowing programmers to specify the implementing code they wish to call upon.[7] 

When Google set out to design its Android operating system for smartphones, it initially sought to license Java SE from Sun, including the Sun Java API.  But these negotiations fell apart, and Google ultimately developed its own platform, which would be compatible with legacy Java applications.[8]  Google wrote millions of lines of new code underlying the platform.  However, Google copied verbatim approximately 11,500 lines of Java SE declaring code for 37 packages that would call upon certain tasks to be executed, even though it wrote the extensive implementing code that would actually carry out the tasks.[9]  Google did this so that developers conversant with the Java language could call upon certain tasks in a manner that was familiar to them.[10]

In 2010, Oracle America, Inc. (“Oracle”) purchased ownership rights to Java and brought suit against Google for patent and copyright infringement, ultimately seeking over $8 billion in copyright damages.  Oracle claimed that Google had violated its copyright in the 11,500 lines of declaring code.  Oracle lost on its patent claims and subsequently abandoned them.  As for copyright infringement, a jury found that Google had incorporated the declaring code into its Android platform, but hung on whether Google’s actions constituted fair use.  However, the district judge then ruled that, as a matter of law, the declaring code was not copyrightable because it was a “system or method of operation,” something excluded from copyright protection by statute.[11]  The Federal Circuit reversed on appeal, holding that the declaring code was copyrightable, and it remanded for another trial on whether Google’s use of the declaring code constituted fair use.[12]

Back in the district court, Google won again—persuading a new jury that its use of the declaring code qualified as fair use.  But the Federal Circuit once again reversed on appeal, stating that “[t]here is nothing fair about taking a copyrighted work verbatim and using it for the same purpose and function as the original in a competing platform.”[13]  The Supreme Court granted Google’s petition for a writ of certiorari on both the copyrightability and fair use issues.[14]

The Supreme Court’s Decision

In its 6–2 majority decision, the Court declined to answer the threshold question of whether the declaring code was copyrightable.  Reasoning that “the rapidly changing technological, economic, and business-related circumstances” cautioned in favor of answering no more than is needed to resolve the parties’ dispute, the Court assumed for the sake of argument that the entire Sun Java API was copyrightable.[15]  In this respect, the Court’s opinion was a disappointment for those eagerly awaiting a ruling setting out the standards for assessing whether particular forms of software qualify for copyright protection—and in particular an answer to the question of whether APIs are entitled to protection. 

Instead, the Court focused on whether Google’s use of the declaring code qualified as fair use, and concluded that it did.  The Court first considered what standard an appellate court should use in reviewing a fair use determination.[16]  The Federal Circuit had concluded that fair use constitutes a mixed question of fact and law, and that although courts should appropriately defer to a jury’s findings of underlying facts, the ultimate question of whether those facts establish fair use is a legal question for judges to decide de novo.[17]  The Supreme Court agreed, observing that fair use was “originally a concept fashioned by judges” and concluding that “the ultimate ‘fair use’ question primarily involves legal work.”[18]  The Court thus rejected Google’s arguments that the Court should take a deferential approach of considering only whether “substantial evidence” supported the jury’s decision that Google’s use of the declaring code constituted fair use, and that to do otherwise runs afoul of the Seventh Amendment.[19]

On the substance of the fair use arguments, the Court observed that “fair use can play an important role in determining the lawful scope of a computer program copyright,” by “providing a context-based check” to “distinguish among technologies” and “keep a copyright monopoly within its lawful bounds.”[20]   

The Court then engaged in an analysis of the four factors that the Copyright Act sets out as non-exclusive considerations in a fair use assessment: the “purpose and character of the use,” the “nature of the copyrighted work,” the “amount and substantiality of the portion used in relation to the copyrighted work as a whole,” and the “effect of the use upon the potential market for or value of the copyrighted work.”[21]  The Court concluded that each factor weighed in favor of a finding of fair use.

The Court began by examining the nature of the copyrighted work at issue.  It characterized the declaring code at issue as distinct from other computer programs because it “is inherently bound together with uncopyrightable ideas (general task division and organization) and new creative expression (Android’s implementing code),” and “its value lies in its efforts to encourage programmers to learn and to use that system so that they will use (and continue to use) Sun-related implementing programs that Google did not copy.”[22]  Accordingly, the Court concluded that “the declaring code is, if copyrightable at all, further than are most computer programs (such as the implementing code) from the core of copyright,” and that finding fair use on the present facts would be unlikely to undermine the general copyright protections that Congress provided for computer programs.[23]

As to the purpose and character of the use, the Court found that even though Google’s use of the code was commercial in nature, this was not dispositive.  The Court gave weight to the fact that Google copied no more than the necessary portions of the Sun Java API in order to “create a new platform,” the Android platform, which is consistent with the “creative ‘progress’ that is the basic constitutional objective of copyright itself.”[24]

Next, with respect to the amount and substantiality of the portion used, the Court noted that the 11,500 lines that were copied amounted to only 0.4 percent of the total code in the Sun Java API,[25] and it gave weight to the fact that the use was “tethered to a valid, and transformative, purpose.”[26]

Finally, as to market effects, the Court cited to evidence that Sun itself “was poorly positioned to succeed in the mobile phone market”—in part because Sun had tried and failed to enter this market—and found that the jury reasonably concluded that Google “did not harm the actual or potential markets for Java SE,” given the distinct markets for computers (in which Sun had succeeded in establishing Java SE) and mobile devices (for which Google had used elements of Java SE).[27]  In doing so, the Court distinguished between value derived from new users being attracted to a program because of its expressive qualities, and a program that is valuable simply because users are accustomed to it.[28]  The value of using the Sun Java API with the Android platform thus had more to do with the investment of third parties (the programmers) in learning and using Sun Java programs than Sun’s investment in creating the Sun Java API.[29]

Based on these factors in combination, the majority of the Court concluded that Google’s use of Sun Java API declaring code qualified as fair use.

In a dissenting opinion, Justice Thomas, joined by Justice Alito, took issue with both the majority’s decision to bypass the copyrightability question and its conclusion that Google’s use of the declaring code constituted fair use.  He argued that declaring code is protected by copyright, reasoning that declaring code is inextricably bound together with implementing code, and that the two must “typically rise and fall together.”[30]  Justice Thomas further asserted that the majority’s decision on fair use “makes it difficult to imagine any circumstance in which declaring code will remain protected by copyright,”[31] and that the majority’s analysis failed to take account of the differences in the Google and Oracle business models (with the latter centered around licensing its intellectual property), as well as the similar ways in which the two companies used Java code.[32]  The majority found these arguments unpersuasive and took the view that the dissent’s “all or nothing” approach to the copyrightability of computer programs would be inconsistent with the overall design of the Copyright Act and the copyright framework for other subjects.[33]

Key Impacts

Applying copyright principles to computer programs has been a notoriously difficult enterprise with inevitable tensions.  On one hand, the Copyright Act expressly provides that a “computer program”—which it defines as “a set of statements or instructions to be used directly or indirectly in a computer in order to bring about a certain result”—can be copyrighted.  17 U.S.C. §§ 101, 109(b), 117, 506(a).  On the other hand, the Act expressly excludes copyright protection for “any idea, procedure, process, system, method of operation, concept, principle, or discovery.”  Id. § 102(b).  As the Court itself observed, “[t]he fact that computer programs are primarily functional makes it difficult to apply traditional copyright concepts.”[34]

At oral argument, several justices seemed to struggle with the first issue on which they had granted review:  whether or not the declaring code at issue should be deemed copyrightable.  If the argument is that the answer should be no, because declaring code is more functional than expressive, how could that line be drawn in a way that would not effectively disqualify most computer programs from protection?  Ultimately, the Court declined to dispose of the case on this binary basis and instead opted to focus on the more nuanced framework for assessing fair use.

The upshot is a highly case-specific decision that is unlikely to directly dictate the result in future disputes.  After all, to be on all fours with this precedent, a future case would need to entail the use of limited portions of a highly popular programming language, valued more for its familiarity than its creative expression, by a company seeking to develop a platform in a new landscape—a field that the entity that created the language had tried and failed to enter.

Still, the decision may have far-reaching implications.  As the Court emphasized, “fair use can play an important role in determining the lawful scope of a computer program copyright.”[35]  And the fact that the Court found fair use even though the defendant used the copyrighted content in a commercial venture, after having tried and failed to negotiate a license that would have encompassed the content, will make the decision a friend of those seeking to invoke the doctrine in similar circumstances.

One context in which the decision may come into play will be the use by a company of another company’s software to achieve interoperability of products.  In making the aforementioned observation that fair use can play an important role in addressing computer program copyrights, the Court cited to two decades-old Ninth Circuit decisions: Sony Computer Entertainment, Inc. v. Connectix Corp.[36] and Sega Enterprises Ltd. v. Accolade, Inc.[37]  In both cases, the defendants had “reverse-engineered” code in order to understand how certain video game consoles worked (a process also referred to as “intermediate copying” of the code).  In Sony Computer Entertainment, the defendant used the knowledge gained from the reverse-engineering process to create an emulator by which Sony PlayStation games could be played on a personal computer; in Sega Enterprises Ltd., the defendant used the knowledge gained from reverse-engineering to create new games that could be played on a Sega Genesis.  In both cases, the defendant’s final product did not contain any infringing code, and in both cases, the court found that the reverse-engineering of the plaintiff’s code qualified as fair use.  Although the Court’s citation to these cases does not, of course, constitute wholesale adoption of their holdings, it suggests an openness to a fair use argument in the context of expanding access to a platform or allowing the use of products on a new platform, provided that the party invoking the defense does not unfairly compete with the owner of the software or infringe other applicable rights (such as trade secrets and patents) or violate license restrictions.

Additionally, as explained above, the Court clarified that although an appellate court should give weight to a jury’s findings of underlying facts, the question of whether those facts establish fair use is a legal question to be decided by judges and reviewed on appeal de novo.  As a result, the contours and applications of the fair use doctrine will not be driven by jury verdicts but instead will be fashioned and policed by judges as a matter of ongoing jurisprudence.


[1] No. 18-956, 2021 WL 1240906 (U.S. Apr. 5, 2021).

[2]  17 U.S.C. § 107.

[3]  Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569 (1994).  The Supreme Court held that a rap parody of Roy Orbison’s rock ballad “Oh, Pretty Woman” was fair use despite the fact that the derivative use was commercial in nature. 

[4] See, e.g., Authors Guild, Inc. v. Google, Inc., 804 F.3d 202 (2d Cir. 2015) (digital copying of millions of books from research library collections constituted fair use in part because the uses were transformative, the display of copyrighted material was properly limited, and the use did not serve as a market substitute for the original works). 

[5] 2021 WL 1240906 at *4–5.

[6] Id. at *4.

[7] Id. at *5.

[8] Id. at *4.

[9] Id. at *6.

[10] In other words, “[b]y using the same declaring code for [the 37] packages, programmers using the Android platform can rely on the method calls that they are already familiar with to call up particular tasks . . . ; but Google’s own implementing programs carry out those tasks.  Without that copying, programmers would need to learn an entirely new system to call up the same tasks.”  Id.

[11] Oracle Am., Inc. v. Google Inc., 872 F. Supp. 2d 974, 1002 (N.D. Cal. 2012).

[12] Oracle Am., Inc. v. Google Inc., 750 F.3d 1339, 1381 (Fed. Cir. 2014).

[13] Oracle Am., Inc. v. Google LLC, 886 F.3d 1179, 1210 (Fed. Cir. 2018).

[14] See 2021 WL 1240906, at *8.

[15] Id. at *10.

[16] Id. at *11–12.

[17] Id. at *11.

[18] Id. at *12.

[19] Id. at *11–12.

[20] Id. at *11.

[21] Id. at *9 (citing 17 U.S.C. § 107).  The Court noted that the factors are not exhaustive and that some factors may prove more important in some contexts than others.  Id.

[22] Id. at *14.

[23] Id.

[24] Id. at *15.

[25] Id. at *16.

[26] Id. at *17.

[27] Id. at *18.

[28] Id. at *19.

[29] Id.

[30] Id. at *23 (Thomas, J., dissenting).

[31] Id.

[32] Id. at *24–29.

[33] Id. at *11.

[34] Id. at *19.

[35] Id.; see also id. (“[Fair use] can help to distinguish among technologies.  It can distinguish between expressive and functional features of computer code where those features are mixed.  It can focus on the legitimate need to provide incentives to produce copyrighted material while examining the extent to which yet further protection creates unrelated or illegitimate harms in other markets or to the development of other products.  In a word, it can carry out its basic purpose of providing a context-based check that can help to keep a copyright monopoly within its lawful bounds.”).

[36] 203 F.3d 596, 603–08 (9th Cir. 2000).

[37] 977 F.2d 1510, 1521–27 (9th Cir. 1992).