Google v. Oracle IV: Fair use & the difference between new and transformative.

Although it has been my intention to write about Google v. Oracle serially, addressing the legal questions in more or less in the order they are presented and weighed in a court opinion, it turns out today marks the end of Fair Use Week.  (How I could have missed that in this otherwise sleepy news cycle is a mystery, I know.) But as Fair Use Week is still officially live, I am going to jump ahead in this post to respond to Google’s claim that its use of Oracle’s Java SE code in the development of the Android platform was a fair use.

We will assume for the sake of discussion that Google’s challenge to the copyrightability of Oracle’s code will not succeed because, absent an infringement, there is no reason to consider a fair use defense. On that note, it is worth mentioning that while it may be good legal strategy to present a fair use argument as a Plan B in a litigation, some fair use assertions are more demonstrably hail-Mary plays than others.  And in this case, Google’s argument seems like a pretty wild pass all the way down the gridiron that should be knocked down by the fair use test.

Above all, Google’s fair use assertion under the first factor—arguing that its use of Oracle’s code was transformative—is yet another example of this tech giant in particular seeking to conflate the novelty of a product with the nature and purpose of transformativeness in fair use. 

For quick review, transformativeness, in its earliest application, tilts toward a finding of fair use when a new creative expression is derived from the specific use at issue.  Hence, the seminal case Campbell  v. Acuff-Rose (1994), in which the Supreme Court unanimously held that 2 Live Crew’s use of the heart of the song “Oh, Pretty Woman” produced a new expression—a parody of the original. Campbell upholds the purpose of copyright to promote new forms of expression such that society gains both the original work and the parodic comment upon the original work. 

By contrast, in considering Google Books, a search tool that relies on the use of digitized copies of millions of published works, the courts in 2015  broadened the doctrine to encompass uses that are transformative because they “expand the utility” of the original material. The Google Books interface offers an unprecedented and highly-useful research tool that does not provide a substitute for the works used—namely it does not make full books under copyright available.

Nevertheless, the Second Circuit Court of Appeals cautioned that its holding in Google Books “tested the boundaries of fair use.”  In other words, the “utility” aspect of the transformative test is meant to be scrutinized very carefully, as the same court later affirmed in its (2018) consideration of the service ReDigi, which asserted that an online exchange trading in “used” digital music files was transformative …

“Even if ReDigi is credited with some faint showing of a transformative purpose, that purpose is overwhelmed by the substantial harm ReDigi inflicts on the value of Plaintiff’s copyrights through its direct competition in the rights holders’ legitimate market, offering consumers a substitute for purchasing from the rights holders.”

Translation:  not everything “new” is transformative under a fair use analysis. Google’s claim that its use was transformative in its defense against Oracle breaks the boundaries the Second Circuit drew in Google Books because Google did nothing to “expand the utility” of Oracle’s code.  On the contrary, Google used Oracle’s code for the exact purpose for which it had been developed—and for which other mobile developers had licensed the work.

Further, Google’s claim is not markedly distinguishable from the holding in ReDigi; Google’s use of Oracle’s code put the search giant in “direct competition” with the party whose work it appropriated, usurping opportunities in the mobile market that Oracle was already exploiting by licensing its products.

Google and supporting amici assert that the roll-out of Android itself is sufficient to render its use of Oracle’s code transformative. But if mere “newness” of a product (or even a creative work) were the shibboleth required to pass the transformative test, this standard would swallow copyright in its entirety and nullify the purpose of a fair use exception.  “… the more amorphous and unreasonably expansive the analysis and application of the fair use doctrine, the harder it becomes to establish the value of the copyrighted work during licensing negotiations that are the lifeblood of the creative ecosystem,” states the brief filed by songwriters in support of Oracle.

Any use of a prior work will always result in something new; but this novelty alone has never relieved the user of the responsibility to either license the prior work or to demonstrate how the use narrowly qualifies for a fair use exception. In this case, Google makes the familiar (though thankfully still unsuccessful) argument whereby the infringing user asserts that migrating a work from one medium to another is transformative. Not only is this not transformative, but in Google’s case, using Java SE in mobile platforms is not even novel. In the absence of transformativeness the first factor consideration of commercial v. non-commercial use will weigh heavily against Google given that Android is a multi-billion-dollar commercial use.

On the third and fourth fair use factors, Google should also fail, while it may end up a draw on the second.  The second factor considers the nature of the work used, and although neither party denies creativity in the declaring code at issue, the inherent functionality of software may point towards a tie in the analysis of the Court. The third factor considers the amount of the original work used, and although Google emphasizes that it copied only a fraction of Java SE, Oracle states in its brief that, “Google admits it copied the packages most valuable to create a derivative version of Java SE for mobile devices.” Further, fair use is not sustained by showing how much they did not copy.

The fourth factor should be especially prejudicial against Google’s fair use claim, as it addresses the harm, and potential harm, to the creator’s market for the work used.  Here, as in ReDigi, the analysis militates against fair use.  As Oracle states in its brief, describing prior licensees of Java…

“If what Google did was permissible, IBM, Danger, and others would not have licensed Sun’s declaring code or complied with [the interoperability standard] ‘write once, run anywhere.’ If everyone could copy the declaring code without a license, Java SE would lose value, as anyone could ‘reimplement’ a knock-off. This undisputed evidence negates Google’s defense as a matter of law.”

More broadly, if Google’s unprecedented assertion of fair use in this case were the new standard, this would only empower the wealthiest corporations to poach any creative works they choose, as long as whatever they use them for has not already been put on the market. That predicate offends the purpose of copyright and is anathema to the interests of all creators in all media. Google enjoys enough advantages when it comes to squashing competitors and making a business out of infringement, without the courts also rewriting decades of copyright doctrine at their behest. 

Enjoy this blog? Please spread the word :)