Repurposed Foul Lines after Google v. Oracle

The majority opinion was somewhat disappointing, at least to some, in avoiding the first of two questions presented by Google: whether application programming interface (API) “declaring code” and its associated organizational structure is copyrightable. The Google decision, however, has significant relevance to copyright law in that it squarely addresses the other question: whether Google’s admitted copying was protected by the fair use defense. Certainly, the decision did not resolve all fair use issues, and it may have created some new questions.² Nevertheless, the decision provides guideposts for the resolution of future copyright fair use disputes over software code, particularly regarding how to determine if the copying of code qualifies as transformative. The Court also provided helpful procedural guidance on how courts should approach resolution of the fair use issue, sorting through the mixed questions of law and fact.

Factual and Procedural Background

Google purchased Android in 2005. It sought to develop a software platform for smartphones that was free and open to software developers. Sun Microsystems had earlier created a software platform that allowed developers using the Java language to write programs that could run on any computer, regardless of the operating system. Google copied the declaring code of 37 Java API “packages” to create the Android smartphone operating system. Oracle acquired Sun in 2010 and sued Google for copyright infringement shortly thereafter. The first jury deadlocked on the fair use issue, and the district court ruled that the API declaring code and the associated “structure, sequence, and organization” (SSO) could not be protected by copyright.³ However, applying Ninth Circuit law, the Federal Circuit reversed and remanded.⁴

The second jury trial resulted in a finding of fair use. The district court denied Oracle’s new trial motions. The Federal Circuit again reversed on appeal.⁵ On appeal, it was undisputed that Google copied some 11,500 lines of the declaring code from Java for 37 API packages, including the names given to particular tasks and the grouping of those tasks into classes and packages. Most relevant to the pivotal transformative issue (discussed below), the parties had stipulated that Google “used the API packages in Android for the same purpose they were created for in Java.”⁶

The Supreme Court granted Google’s certiorari petition. It issued the Google decision on April 5, 2021. The Court reversed the Federal Circuit’s decision on the fair use defense and reinstated the district court’s judgment in favor of Google on that ground. That ruling has important implications for the path of future litigation over the fair use defense.

Copyrightability Was Assumed but Not Decided

Some had hoped the Supreme Court would conclusively answer the question of whether the declaring code of an API and its associated SSO are protected by copyright. The 6–2 majority (Justice Barrett did not participate) chose not to address that threshold question. Its opinion, authored by Justice Breyer, cited the “rapidly changing technological, economic, and business-related circumstances” as the reason not to answer “more than is necessary” to resolve the dispute before it.⁷ “[P]urely for argument’s sake,” it thus assumed that the declaring code in question “falls within the definition of that which can be copyrighted.”⁸ This was also the conclusion reached by the Federal Circuit on the first appeal (reversing the district court) and the dissent (written by Justice Thomas and joined by Justice Alito).

While avoiding the copyrightability issue, the Google decision is undeniably an important precedent regarding the scope of copyright protection of computer code. In particular, the Court shed light on what is a “transformative” work in the context of the admittedly largely functional computer code used in APIs and how courts should approach the mixed question of law and fact raised by the assertion of the fair use defense, which should have more general applicability to procedure in all types of copyright litigation.

Repurposing to Create a New Product: Fair and Transformative Use

Writing for “lay persons,” including judges and juries, the Court provided a short tutorial on what an API is and its importance to programmers.⁹ It then summarized the “complex and lengthy history” of the 10-year-old litigation.¹⁰ Next, it turned to the fair use issue.

Undergirding the Court’s analysis is the constitutional purpose behind statutory copyright protection—the need “[t]o promote the Progress of Science and useful Arts.”¹¹ The Court explained that the copyright statute grants exclusive rights to authors “not as a special reward, but in order to encourage the production of works that others might reproduce more cheaply.”¹² Turning to the work in question (the API declaring code), the Court noted that in 1980 Congress added “computer program[s]” to the definitional list of works eligible for protection under the Copyright Act.¹³ Although it is now codified as 35 U.S.C. § 107, the Court noted that the fair use doctrine originated in the courts as an “equitable rule of reason.”¹⁴

In § 107, Congress sought to codify the common law. The statute lists four factors for consideration by the court. But this listing is “not exhaustive,” and the examples of potentially fair use in the preamble do not prevent other uses from qualifying as a fair use.¹⁵ For “expository purposes,” the Court analyzed the statutory factors out of order.¹⁶

The Court first looked at the second factor, the “nature of the copyrighted work,” before turning to the first, the “purpose and character” of Google’s use. The copyrighted work, the Java API code and its SSO, was a user interface. It provided a means for users to manipulate and control task—or method—performing programs via menu commands. Because of its largely functional purpose, the copyrighted API declaring code and its SSO was further away from the core of copyright than other computer programs. As a result, this factor pointed to thin copyright protection and in the direction of fair use.

Turning to the first § 107 factor, Google’s use, the Court stated that Google “precisely” copied portions of the API declaring code “in part for the same reason that Sun created them”: to allow programmers to call up the implementing code for particular tasks.¹⁷ However, the analysis cannot stop with that functional (and largely uncopyrightable) feature, as computer code is always functional. To determine “whether a use is ‘transformative,’ we must go further and examine the copying’s more specifically described ‘purpose[s]’ and ‘character.’ 17 U.S.C. § 107(1).”¹⁸ The Court placed considerable weight on the purpose behind Google’s use of the API “to expand the use and usefulness of Android-based smartphones,” thereby creating a “new product [that] offers programmers a highly creative and innovative tool.”¹⁹ Again, quoting Article I of the U.S. Constitution, the Court described this as the “basic” and “primary” objective of “copyright itself.”²⁰

Both the Federal Circuit and the dissent looked at these same facts differently. They emphasized the fact—conceded by the majority—that Google used the API for the “same purpose” as did the author, Sun: to enable programmers to call up the implementing code for particular predetermined tasks. They argued that “a use becomes transformative only if it serves a different purpose.”²¹

The fundamental difference of opinion is whether or not a copyist’s use of thinly protected computer code to create “new products” is transformative.²² Stated differently, the Google Court concluded that “reimplementing” or “repurposing” the copied portions of the API for use on a smartphone meant that the “purpose and character” of Google’s use was transformative. The Court noted the various amicus briefs asserting that the largely functional copyrighted API code could become an industry standard and, unless restricted, would give the copyright owner anticompetitive power. Allowing reimplementation of that code on another platform would fuel widespread adoption, enable innovation, and foster growth of the entire market.²³

Although the Court reviewed all four fair use factors under § 107, the dispositive issue turned on whether or not Google’s admittedly commercial use of the declaring code qualified as transformative. As Google used the copied parts of the API to create a new platform that could be readily used by programmers to create new works, the Court held that its use was consistent with that creative “progress” that is the basic constitutional objective of copyright. These and related facts convinced the Court that the “purpose and character” of the use by Google was “transformative.”²⁴ The Court acknowledged that § 107 lists various noncommercial uses as “paradigmatic” examples of fair use and that, in contrast, “Google’s use was a commercial endeavor.”²⁵ It held, nevertheless, that this was not dispositive in light of the “inherently transformative role” that the use played in the “reimplementation” of the copied code.²⁶

Having found the copying transformative, the Court applied that conclusion to find that the last two factors also favored fair use. Although the 11,500 lines copied might at first blush seem substantial enough to tilt the third factor—“amount and substantiality of the portion used”—against Google, the Court pointed out that this is only 0.4% of the 2.86 million lines of code in the Sun Java API code.²⁷ While even a small amount of copying at the “heart” of the protected expression might be “substantial,” the better way to look at the copying here is to consider that Google did not copy several million lines of the associated task-implementing API code. Google copied only the declaring code for its function, to call up the task-implementing code (which Google created on its own). Programmers knew the Sun API and, the Court reasoned, “it would have been difficult, perhaps prohibitively so, to attract programmers to build its Android smartphone system without” using the Sun API declaring code.²⁸ The Court concluded that “[t]he ‘substantiality’ factor will generally weigh in favor of fair use where, as here, the amount of copying was tethered to a valid, and transformative, purpose.”²⁹

Finally, as to the fourth factor—“market effects”—the transformation into a different product led logically to the conclusion that the API used by programmers for the purpose of creating smartphone apps is distinct from software used on computers. In particular, the Court decided that the jury could have found that Google’s development of API for Android “did not harm the actual or potential markets for Java SE.”³⁰ The Court was careful to keep its holding narrow, noting that it was not saying this was “always relevant to the application of fair use” of computer programs.³¹ The Court did acknowledge (as the Federal Circuit and dissent emphasized) that Java code had been used in other mobile phones. That business, the Court said, was declining, and there was market demand for “a new form of smartphone technology that Sun was never able to offer.”³² The jury might also have found that “Android and Java SE [are] operating in two distinct markets.”³³ Lastly, the Court reasoned that giving Oracle the sole right to use the API code to develop new creative technology in smartphones would potentially lock out the public, which would “interfere with, not further, copyright’s basic creativity objectives.”³⁴

Litigation over What Is a Transformative Use of Computer Code

There are thousands of open APIs available and routinely used by third-party programmers. Programmers, or more likely their lawyers, who strive to provide guidance to minimize the risk of copyright infringement, may see the Google decision as expanding the fair use protection. The precedent should at least protect that status quo and perhaps encourage more extensive copying of API code. Greater unauthorized use of APIs may foster more litigation, more reliance on the fair use defense, and likely more conflicts over how to apply the fair use factors in light of the Google Court’s teachings.

Pivotal to the Court’s analysis was the determination that Google copied and used the API declaring code to create new products. To the extent that Google used parts of the Java API to create a new platform that could be readily used by programmers, the Court reasoned, its use was consistent with creative “progress,” a basic constitutional objective of copyright. Further, the Court emphasized that Google copied only the API needed to allow programmers to call upon tasks using the familiar Java programming language without having to learn a new one. In the Android platform, those tasks were carried out by new implementing code (written by Google). Given these broad policy rationales supporting the Google Court’s decision, disputes over whether certain instances of copying API code are transformative seem likely, if not inevitable.

Some may argue, for example, that the Court’s holding that copying the declaring code of an API is fair use supports the conclusion that copying the implementing code of an API should also be fair when used to create “new products.” At least some of the same policy issues could be argued to apply to API implementing code. Indeed, the same could be said more generally—that allowing even greater “fair use” use of all API code should promote even more rapid development of new products and would thus also be “transformative.” API software authors may respond to this threat by keeping their code secret by requiring nondisclosure agreements to access it. That, too, may lead to future conflicts.

Another unanswered issue is where to draw the line between a derivative work (protected by copyright) and the transformational work or “new product” (not protected by that copyright) that can be a fair use. The Google Court recognized that a copyright on software also protects derivative works but did not discuss the issue further.³⁵ The dissent accused the majority of “wrongly conflat[ing] transformative use with derivative use.”³⁶ It seems likely that this issue, too, will be raised and debated in future disputes over the scope of the fair use defense.³⁷

Fair Use May Be Ripe for Decision on Summary Judgment

The Court held that Google’s unauthorized use of those thousands of lines of declaring code in its Android operating system was “fair” as a matter of law. The jury’s ultimate conclusion on fair use was necessarily only advisory; fair use is a question of law for a judge rather than a jury.

The Federal Circuit had interpreted “governing Supreme Court and Ninth Circuit case law” as requiring that “[a]ll jury findings relating to fair use other than its implied findings of historical fact” are “advisory only.”³⁸ The Federal Circuit’s decision in Oracle II had noted that as of the time of its writing, “the Supreme Court has never clarified whether and to what extent the jury is to play a role in the fair use analysis.”³⁹ The Google decision now provides that Supreme Court clarification. The Court agreed with the Federal Circuit that “the ultimate question whether [the underlying] facts showed a ‘fair use’ is a legal question for judges to decide de novo.”⁴⁰ Applying its recent U.S. Bank decision, the Court held that “the ultimate ‘fair use’ question primarily involves legal work.”⁴¹

On this issue, the Supreme Court rejected Google’s argument that the issue must go to the jury, and it endorsed the Federal Circuit’s approach, which had applied Ninth Circuit law.⁴² The Federal Circuit characterized this approach as the “modern view” and rejected the view in earlier cases suggesting that fair use is normally a jury question.⁴³ This is an important clarification of the law that appears to have general application to fair use defense in copyright cases. Litigants may thus now rely on this portion of the Federal Circuit decision as good law, endorsed by the Supreme Court, though the decision itself was “reversed on other grounds.” Armed with these two decisions, courts and parties have a better foundation on which to build upon the well-established summary judgment framework for analysis.

The party asserting the fair use defense has the burden of proof. To warrant summary judgment, it must provide evidence that would be sufficient to support a directed verdict in its favor.⁴⁴ Although proper inferences to be drawn from the four-factor § 107 analysis are legal questions, “in the Ninth Circuit, disputed historical facts represent questions for the jury.”⁴⁵ Here, there is a distinction between the facts themselves, such as the history, content, and origin of the copyrighted work, and the who, what, where, and when of the allegedly infringing use of the work.⁴⁶ These facts may not be genuinely in dispute. Indeed, Google and Oracle had stipulated to many of these facts, including that Google’s use was commercial. Thus, the question of whether the fair use defense is ripe for decision on summary judgment will often turn on whether the underlying historical facts—not the proper inferences or conclusions to be drawn from them—are genuinely in dispute.⁴⁷

The Future Impact of Google on Fair Use Case Law

Consistent with the prevailing approach of the Roberts Court, the Google Court believed it “should not answer more than is necessary to resolve the parties’ dispute.”⁴⁸ Several times it emphasized the narrowness of its holding as confined to the difficult area of computer programs that are primarily functional.⁴⁹ The Court took “the principles set forth in the fair use statute, § 107, and set forth in our earlier cases,” involving written works and movies, and “applied them to this different kind of copyrighted work,” computer code.⁵⁰

While narrow, the Google decision is undeniably an important copyright precedent. It is premature to predict the significance of the Google decision on yet-to-be-developed fair use case law outside of the computer field. Litigants, however, may rely on the decision to argue that (1) copying API code is an efficient means to create new products and is therefore a “transformative” use potentially eligible for fair use protection; and (2) the fair use issue is a legal issue ripe for summary judgment in many cases where the underlying historical facts cannot reasonably be disputed.

Endnotes

1. 141 S. Ct. 1183 (2021).

2. In response to a petition for rehearing relying on the Google decision, the Second Circuit recently issued a revised decision in the Warhol case on fair use. Andy Warhol Found. for the Visual Arts, Inc. v. Goldsmith, 11 F.4th 26 (2d Cir. 2021). The court granted rehearing “to consider fully the Supreme Court’s most recent teaching on fair use” and issued an amended decision. While it did not change the outcome—summary judgment of no fair use—references to the Google decision are sprinkled in throughout along with seemingly nonsubstantive deletions and alterations to the text.

3. See Oracle Am., Inc. v. Google Inc., 872 F. Supp. 2d 974, 999–1000 (N.D. Cal. 2012).

4. Oracle Am., Inc. v. Google Inc., 750 F.3d 1339, 1348 (Fed. Cir. 2014).

5. Oracle Am., Inc. v. Google LLC (Oracle II), 886 F.3d 1179 (Fed. Cir. 2018).

6. Id. at 1196.

7. Google LLC v. Oracle Am., Inc., 141 S. Ct. 1183, 1197 (2021).

8. Id.

9. See id. at 1191–94.

10. See id. at 1194–95.

11. Id. at 1195 (quoting U.S. Const. art. I, § 8, cl. 8).

12. Id.

13. Id. at 1196.

14. Id. (quoting Stewart v. Abend, 495 U.S. 207, 236 (1990)).

15. Id. at 1197. The Court quoted most of § 107, which it said only “indicates, rather than dictates” how courts should apply the defense:

[T]he fair use of a copyrighted work . . . for purposes such as criticism, comment, news reporting, teaching . . . scholarship, or research, is not an infringement of copyright. In determining whether the use made of a work in any particular case is a fair use the factors to be considered shall include—

(1) the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;

(2) the nature of the copyrighted work;

(3) the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and

(4) the effect of the use upon the potential market for or value of the copyrighted work.

Id. at 1196–97 (quoting 35 U.S.C. § 107).

16. Id. at 1201. Justice Thomas’s dissent noted and criticized the majority opinion for this switch in sequence: “Tellingly, the majority evaluates the factors neither in sequential order nor in order of importance . . . . It proceeds in this manner in order to create a distinction between declaring and implementing code that renders the former less worthy of protection than the latter.” Id. at 1214–15 (Thomas, J., dissenting). It is true that the majority placed some weight on its conclusion that declaring code is less creative and thus further from core copyright protection than implementing code. The dissent, however, concedes that computer code by its nature is predominantly functional and focuses on the “value” of the two types of code (which it asserts is comparable) rather than the creativity required to create them.

17. Id. at 1203 (majority opinion).

18. Id. (alteration in original).

19. Id.

20. Id.

21. Oracle II, 886 F.3d 1179, 1201 (Fed. Cir. 2018); see Google, 141 S. Ct. at 1219 (Thomas, J., dissenting).

22. Compare Google, 141 S. Ct. at 1203 (majority opinion), with id. at 1219 (Thomas, J., dissenting).

23. Google, 141 S. Ct. at 1204 (majority opinion).

24. Id.

25. Id.

26. Id.

27. Id. at 1204–05.

28. Id. at 1205.

29. Id.

30. Id. at 1206.

31. Id.

32. Id. at 1207.

33. Id.

34. Id. at 1208.

35. See id. at 1199.

36. Id. at 1219 (Thomas, J., dissenting) (arguing that a work that serves the “same purpose” as the original in a new context “is derivative, not transformative”).

37. See generally Andy Warhol Found. for the Visual Arts, Inc. v. Goldsmith, 11 F.4th 26, 38–43 (2d Cir. 2021) (discussing Google and noting the “inherent tension in the Copyright Act” between derivative works and transformative fair uses, and concluding that the images at issue there “are closer to what the law deems ‘derivative’ than ‘transformative’”).

38. Oracle II, 886 F.3d 1179, 1194 (Fed. Cir. 2018).

39. Id.

40. Google, 141 S. Ct. at 1199.

41. Id. (citing U.S. Bank N.A. v. Village at Lakeridge, LLC, 138 S. Ct. 960, 967 (2018)).

42. The Federal Circuit surveyed Ninth Circuit law, finding cases that sometimes submitted the fair use issue to the jury without “analysis of why it may.” Oracle II, 886 F.3d at 1194. It concluded: “[W]hile inferences from the four-factor analysis and the ultimate question of fair use are ‘legal in nature,’ in the Ninth Circuit, disputed historical facts represent questions for the jury. Where there are no disputed material historical facts, fair use can be decided by the court alone.” Id. at 1195 (citation omitted).

43. Id. at 1194 n.3. In its footnote, the court noted the seminal case on fair use, written by Justice Story, that “described fair use as a ‘question of fact to come to a jury’ in 1845. Emerson v. Davies, 8 F. Cas. 615, 623–24 (C.C.D. Mass. 1845).” The Federal Circuit also cited a modern example of this older, disapproved view in DC Comics, Inc. v. Reel Fantasy, Inc., 696 F.2d 24, 28 (2d Cir. 1982), where the court “found that ‘[t]he four factors listed in Section 107 raise essentially factual issues and, as the district court correctly noted, are normally questions for the jury.’”

44. See C.A.R. Transp. Brokerage Co. v. Darden Rests., Inc., 213 F.3d 474, 480 (9th Cir. 2000) (“When the party moving for summary judgment would bear the burden of proof at trial, ‘it must come forward with evidence which would entitle it to a directed verdict if the evidence went uncontroverted at trial.’” (quoting Houghton v. South, 965 F.2d 1532, 1536 (9th Cir. 1992))).

45. See Oracle II, 886 F.3d at 1195 (citing cases).

46. Id. at 1194.

47. Id. at 1195 (“Where there are no disputed material historical facts, fair use can be decided by the court alone.”).

48. Google LLC v. Oracle Am., Inc., 141 S. Ct. 1183, 1197 (2021).

49. Id. at 1208 (“We do not overturn or modify our earlier cases involving fair use—cases, for example, that involve ‘knockoff’ products, journalistic writings, and parodies.”).

50. Id. at 1208–09.