The U.S. Court of Appeals for the Federal Circuit considered whether Google’s copying of Oracle’s Java API for integration into its Android smartphone platform was fair use. On March 27, 2018, the Court held that it was not, and reversed the lower court’s determination notwithstanding a favorable jury verdict for Google on fair use.

Before considering the fair use factors, the Court recounted the various technical details of the Java API and Google’s Android platform, as well as the various determinations, appeals and remands in the case. The Court also delved into the appropriate standard of review, noting that fair use is mixed question of law and fact, and recounted the U.S. Supreme Court’s recent decision in U.S. Bank Nat’l Ass’n ex rel. CWCapital Asset Mgmt. LLC, No. 15-1509, 2018 WL 1143822, at *5 (U.S. Mar. 5, 2018). This decision held in part that where issues “utterly resist generalization” they should be treated as factual issues and accorded deference.

Although many of the facts here may defy generalization, the Court found that modern Ninth Circuit precedent treated fair use as an “equitable rule of reason,” and consequently, a legal question subject to de novo review, rejecting less “modern” U.S. Supreme Court and Second Circuit precedent that treated fair use as a question of fact. The Court reasoned that all fair use cases are “generally similar” and that assessing fair use in one case “guide[s] resolution of that question in all future cases.” It also found that while fair use is an equitable affirmative defense in the Ninth Circuit, Ninth Circuit courts routinely allow the fair use question to go to a jury, leading the Court to conclude that “in the Ninth Circuit, [only] disputed historical facts represent questions for the jury,” and that a judge must decide the rest.

Despite this conclusion, the Court held that the jury had in fact been given the entirety of the fair use determination, but that it was not reversible error under Ninth Circuit law. Thus, the Court found any portion of the fair use jury verdict that went beyond questions of material historical fact were necessarily “advisory,” that all jury findings on material historical facts were to be given deference, and that because the district court’s jury form did not itemize factual findings, the parties would need to identify the disputed material historical facts for the Court, which were:

  1. whether the use was commercial in nature;
  2. whether Google acted in bad faith in copying the work;
  3. whether there are functional aspects to the copyrighted work that make it less deserving of protection; and
  4. whether there was harm to the actual or potential markets for the copyrighted work.

Turning to the merits, the Court first considered the nature and character of the use, concluding that Google’s use was commercial and that the jury’s finding to the contrary was not supported by substantial evidence. Specifically, the Court rejected the argument that the use was not commercial because Android was given out for free on smartphones, relying on A&M Records, Inc. v. Napster, Inc., 239 F.3d 1004, 1015 (9th Cir. 2001) for the propositions that free goods can still support commercial use and that commercial use does not require that the user receive a direct economic benefit. The Court also noted that partially-non-commercial motivations for the use are irrelevant; rather, the question is whether the user stands to profit from the use without paying the customary price.

The Court then considered the transformative nature of Google’s use, finding it was not transformative. The Court enumerated four reasons for this holding: first, the use was not among those listed in the preamble to 17 U.S.C. § 107; second, the purpose of Google’s use of the API packages was the same as Oracle’s use (to use the Java language); third, Google did not alter the content or message of the copyrighted material; and fourth, Google’s use of the API packages for smartphones did not constitute use in a new context.

In so finding, the Court distinguished Sony Computer Entertainment, Inc. v. Connectix Corp., 203 F.3d 596 (9th Cir. 2000), noting that Connectix’s “modestly transformative” use (a new product using new code that allowed console games to be played on computers) was more transformative than Google’s, which merely used the same code to attract software developers. Nor was the Court swayed by the fact that Google wrote its own implementing code, finding that “[t]he relevant question is whether Google altered ‘the expressive content or message of the original work’ that it copied—not whether it rewrote the portions it did not copy.”

The Court also rejected Google’s chief argument, namely, that use of the Java API on its smartphones was transformative. First, the Court noted that Google was not the first to do this (and referenced Oracle’s licensing activities with Nokia and others). Second, the Court found that moving material to a new medium is not transformative. In so finding, the Court discussed Perfect 10, Inc. v. Amazon.com, Inc., 508 F.3d 1146, 1165 (9th Cir. 2007) to illustrate how changes in medium or context (in Perfect 10, moving the images to the reference tool) are irrelevant, and even an inevitable part of appropriation, whereas changes in purpose (in Perfect 10, repurposing the images into hyperlinks) can be transformative. The Court also briefly considered bad faith, declining to disturb the jury’s apparent conclusion that Google did not act in bad faith, but noting that good faith did not support fair use. In sum, the Court found the first factor weighed against fair use.

Turning to the second factor, the nature of the copyrighted work, the Court found that this factor favored Google because there was substantial evidence for both conclusions. Specifically, there was credible testimony that the Java API was a creative endeavor, and thus subject to strong protection, but also credible testimony that the Java API was primarily functional, and thus subject to lesser protection. Accordingly, the Court was constrained, based on the jury verdict, to find that this factor favored Google, but also noted that this factor was not “terribly significant.”

In considering the third fair use factor, the amount and substantiality of the use, the Court eschewed a strict quantitative approach, noting that such factor “looks to the quantitative amount and qualitative value of the original work used in relation to the justification for its use.” Reviewing the record, the Court found that it was reversible error to conclude that Google only used what it needed from the Java API, as the record indicated that Google used 11,330 lines more than the 170 lines necessary to write in Java. The Court also dispelled the notion that Google used what it needed to achieve Java interoperability, a central theme earlier in the case, noting that Google had jettisoned such arguments after it was revealed that the Android platform was made to be incompatible with Java. Ultimately, the Court found this factor neutral but, if anything, weighed slightly against fair use.

Turning to the fourth factor, the market impact of the use, the Court first noted the relationship between the first and fourth fair use factors, i.e. that non-transformative works are more likely to be market substitutes, and also noted that harm to the “market for potential derivative uses” was properly considered within the fourth factor. The Court pointed to direct competition between Oracle and Google in the smartphone market, namely, four different smartphone-maker Oracle licensees that competed directly with Android phones. As to the tablet market, the Court relied on Amazon’s decision to drop Java SE from its Kindle following Android’s release, and highlighted that “Amazon later used the fact that Android was free to negotiate a steep discount to use Java SE.” Thus, the Court found that Google’s use severely impacted Oracle’s licensing market and/or potential licensing market, rejecting Google’s arguments that Oracle had no smartphone, no platform, and no intentions to develop either, as Oracle’s clear interest in licensing Java for uses like Google’s was sufficient to show market impact.

Balancing the factors, the Court concluded that Google’s use was not fair, and that rote copying of the original code for substantially the same purpose as the original essentially superseded the original, and could not be fair. Reversing the lower court’s decision, the Court remanded for a trial on damages.