April 12, 2021
Following multiple jury trials and two Federal Circuit appeals, the U.S. Supreme Court issued its decision in Google LLC v. Oracle America, Inc. on April 5. The decision addresses important copyright issues related to fair use of computer source code, but does not address the prerequisite issue of whether the code at issue was copyrightable.
In a 6 to 2 decision authored by Justice Breyer, the Court held that Google’s copying of the Sun Java Application Programming Interface (API), owned by Oracle, was a fair use of that material as a matter of law. Google’s copying was therefore defensible, putting an end to Oracle’s claim that it was entitled to almost $9 billion in damages. The Court did not rule on whether the “declaring code” and organizational structure of the Sun Java API are protectible under copyright law, instead assuming “purely for argument’s sake, that the entire Sun Java API falls within the definition of that which can be copyrighted.”
The Court’s decision provides extensive background of the accused copyright infringement. In about 2005, Google started to work on an API for its Android software. Its plans were to use the Java software language that was popular with many developers. In its development of this platform, Google had about 100 engineers spend three years to develop millions of lines of new computer code. However, because Google wanted this software available to as many software developers as possible, it also copied about 11,500 lines of code from the application programming interface of the Java SE software owned by Oracle. Google released its platform in 2007, and within five years it claimed a significant part of the smartphone market. By 2015, its Android sales accounted for more than $42 billion in revenue.
Google’s accused Android project uses the Java API as a user interface that allows programmers to write programs for the Android mobile devices. The technology involves three separate essential parts. The first is implementing code that instructs computers on how to perform a specific task. Google created its own implementing code for the accused program. The second is the method call that instructs computer to run the series of steps. Oracle did not accuse the method call portion of the Google software to infringe its work. Third is the declaring code that was the crux of the infringement claim. The declaring code labels and organizes the tasks for the software to perform.
The initial case was filed in 2010 and, as the court notes, has had a “complex and lengthy history.” That history has been outlined in previous alerts found here and here. The issues accepted for review by the Supreme Court were (1) whether the copied software code is entitled to copyright protection and (2) if so, if the copied use was a “fair use.” Because fair use would moot the issue of the fundamental copyrightability of the work, the court only addressed the issue of fair use, opting to assume “purely for the sake of argument” that the code was copyrightable.
As a procedural matter, the Court addressed the standard of review of the fair use defense, holding that fair use is an equitable defense that is a mixed question of law and fact for which the court is entitled to a de novo review. Once the appropriate factual determinations are made by the jury, it is up to the court to make the ultimate legal determination regarding the application of the fair use defense to those facts.
Under section107 of the Copyright Act, the fair use defense includes an examination of four factors: (1) the purpose and character of the use, (2) the nature of the copyrighted work, (3) the amount and substantiality of the portion of the work used, and (4) the effect of that use on the potential market for the original work. The Court noted that this list is not exhaustive and that some factors may be entitled to more weight than others. The Court’s analysis addressed each of these factors.
The Court analyzed the nature of the copyrighted work first, placing great emphasis on this factor. It concluded that declaring code is further than most computer programs “from the core of copyright.” The Court found this part of the software to be functional in nature and “inherently bound together with uncopyrightable ideas.” These comments question the entire copyrightability of the declaring code, although the Court did not issue a holding on that point.
The Court then analyzed the remaining fair use factors in sequence, starting with purpose and character of the use. This factor includes analysis of whether the accused work adds something new and important that makes the work transformative from the original. Here Google precisely copied portions of the work. But Google sought to use this material to create a new product that expands the purpose and usefulness of the software for the Android smartphone application. Even though it precisely copied portions of the software (which is often the case with software copyright issues), Google only copied that portion of the software that was necessary for it to create the new product. This is sometimes referred to as reimplementation, which is considered necessary in the software industry so that different programs can communicate with each other and so that programmers can take advantage of their existing knowledge of the language. The Court concluded that Google’s software was transformative, which supports the application of the fair use defense.
Other considerations in the purpose and character of the use include commerciality and bad faith. As to the commercial nature of the use of the work, the Court found that a noncommercial purpose usually weighs in support of the fair use defense but the opposite is not automatically the case. Here Google’s use was admittedly commercial in nature. However, just as news reporting—which is expressly entitled to the fair use defense—is usually commercial, the transformative nature of the use here mitigates the commerciality such that fair use can still apply.
As for the issue of bad faith, the Court expressed skepticism as to its usefulness as a factor in this enquiry. Rather than fully explore this factor, the Court held that the strength of the transformative nature of the use outweighs any consideration of good or bad faith in the use.
The third fair use factor is the amount and substantiality of the portion of the work that is used in the accused product. Google directly copied 11,500 lines of code from the Java API declaring code. This is the majority of the declaring code in that API. The entire Java API, however, consisted of 2.86 million lines of code, meaning the copied code was only about 0.4 percent of the software. The Court’s analysis of this factor balanced the amount of copying in view of the nature of the “creative expression” found in the work. For example, copying a minor amount of the whole work can still capture the heart of the work, which would weigh against the fair use defense. Conversely, copying a substantial portion of the work can weigh in favor of the defense if the portion copied “captures little of the creative expression or is central to the copier’s valid purpose.”
Rather than the amount of the work copied, the Court focused on the amount of the work not copied. The Court noted that Google’s purpose was to create an entirely new task-based product for a different environment. According to the Court this is a valid and transformative purpose that tilts the “substantiality” factor in favor of finding fair use. The Court expressly disagreed with the Federal Circuit, reasoning that Google’s valid purpose could not have been accomplished by copying a lesser amount of the work or by creating its own work. As a result, the Court found this factor to also weigh in favor of support of the application of the fair use defense.
The last of the four fair use factors is the effect of the product to the market or the value of the copied work. In addition to loss of revenue, the Court added consideration of the public benefit that the copying is likely to produce. The evidence here established that the primary market for the Java software was desktop and laptop computers, which is a very different market than the mobile phone market. The Court noted that Oracle’s efforts to go into the mobile phone market were largely ineffective. Again, the Court found that there is potentially a public harm by enforcing the copyright against Google’s use, which allows programmers to benefit from their existing knowledge of the software’s organization. As a result, this factor also weighs in favor of the fair use defense in these circumstances.
The Court did not hold that software, or even the “declaring code” at issue, is uncopyrightable. Obtaining copyright protection for computer software consequently still has significant value. Instead, the Court applied fair use broader than many courts had previously allowed. Since much of the Court’s opinion is influenced by arguments that the defendant only used as much of the copyright protected work as was necessary to create an entirely new purpose, potential defendants have new guidance on the use of copyrighted works for new purposes. For the fair use defense, the nature of copied software and the transformative nature of the accused work may become more important. This decision is particularly instructive for computer software copyright claims, and the Court notes that software claims often entail a different analysis than is applied to other, more traditional, copyright claims. Nonetheless, the principles here may be instructive in other copyright matters if the copied work is bound up with uncopyrightable ideas.
For more information on this case and its effects, please contact Fitch Even partner Joseph T. Nabor, author of this alert.