The doctrine of fair use allows people in certain instances to copy or use the work or content of others. The ease of duplication nowadays has caused an increase of copying, and an increase in copyright lawsuits and the defense of fair use. What is fair use? When can one fairly and freely copy something? And what are the constraints on copying?
Fair use is also defined clearly as a limited exception to the various exclusive rights of copyright holders, and allows copying if it is “for purposes such as criticism, comment, news reporting, teaching…, scholarship or research.” 17 U.S.C. 107. Although not an exhaustive list, it is apparent that these uses are usually not commercial, i.e., the use of another’s content to make money. Thus, a professor may employ an image in the context of a lecture, without first obtaining permission of the owner of the image, so long as the use is educational in nature. Similarly, news reporters can employ the words and images of others in writing news articles. In general, so long as no money or little money is involved, Congress has delineated the fair use doctrine as a safe harbor for those legitimately taking the content of others.
Thus, one can assume that the taking by a competitor for a commercial use is not ok. Generally, this is true, however, there are situations where this generality can break down. For example, an artist may parody a work by transforming the original into a new expression, and sell that work. The original owner may disagree on this, but the Supreme Court has indicated that “transformative works,” even if commercial in nature, may be subject to fair use, i.e., there may be a fair use defense to a charge of copyright infringement for some people making money from another’s content.
These issues of fairness and fair use are played out in the recent Oracle v. Google decision. In a convoluted case that has gone up to the Supreme Court once and will again, the Federal Circuit finally was able to make a ruling that the blatant, verbatim copying of computer code is not a fair use. At issue were the copying of 37 Oracle programs or apps, constituting over 11,500 lines of code, by Google for their use in the Android operating system for smart phones and other uses.
It was undisputed that Google did the copying, and two California juries agreed. Google did so after their licensing negotiations with Oracle for the right to use JAVA code broke down ten years ago. When the actions of Google become known to Oracle, they sued for violation of copyright.
A chief issue in the case was whether or not the structure, sequence and organization (SSO) and other code were per se eligible for copyright protection. Despite the creative arguments made that “structured” code cannot be copyrighted, the Federal Circuit disagreed. Indeed, software is a written expression whether structured or not. The intent of this argument was that functional code somehow lost the eligibility for copyright. It is curious that in connection with this case the argument was no copyrighting structured code since you can patent. Yet, software patenting has been condemned since software can be copyrighted! The Federal Circuit stated the obvious that the many tens of millions of lines of JAVA code were all entitled to copyright, including the portion taken.
The effect of this copying was devastating to Oracle, who lost licenses with their customers to Google’s competing platform, and Oracle was also forced to renegotiate other licenses downward in view of this new competition. Clearly, with the huge success of Android, whatever market share Oracle had dwindled, even though JAVA was a popular platform. All of this was due to the earlier corporate decision by Google, after their licensing impasse with Oracle, to copy “JAVA anyway.” Apparently, the tech giant was unable to duplicate the roughly 11,500 lines of proprietary JAVA code. This calculated decision netted Google over $42 billion dollars in ad revenues over the years since. In patent litigation there is a term, “efficient infringement,” where an infringer can just take the patent owner’s technology under the risk of lawsuit, when the upside profit is good enough. Apart from this Federal Circuit decision, this taking constitutes one of the most efficient infringements ever, unless the Supreme Court later decides otherwise. This case will most definitely be appealed.
Another issue in the case involves the interpretation of the term “transformative” in a fair use context, and the effect of this concept in the fair use analysis. There are four factors to be considered in a fair use determination, the first being the purpose and character of the use. If commercial, i.e., the user is making money from the owner’s content, then that weighs heavily against fair use. Under Supreme Court law, however, if a work has been sufficiently transformed, then the commerciality of the use weighs less. Many ingenious arguments were employed by Google’s counsel to show that their taking was transformative.
The Federal Circuit, however, discounted the numerous nuances and went to the heart of what a “transformative use” entailed. The court asked if something new was added, is the work different in character, does it supersede the original, and other questions. The answer was no. The Oracle JAVA code was not altered, but copied verbatim, did the same functions, had the same purposes, was competing, and was thus both “highly commercial” and “non-transformative.”
The second fair use factor concerns the nature of the work itself, e.g., is the work creative in nature, which requires strong copyright protections, or more factual, which is entitled to less or no such protections. Testimonies at trial went both ways. A Google expert, for example, said that these code designs were more “art, than a science,” while others argued that the designs were more functional. The juries, apparently, saw the JAVA code as more functional than creative, rendering this second factor more in favor of fair use. However, this factor is usually the least important.
Turning now to the third factor, the amount and substantiality of the taking, the court held that this was a “wholesale copying” that “militates against a finding of fair use.” An argument, however, was made that even though the entire work was copied, Google took “no more than was necessary,” leaving open a justification defense. However, the Federal Circuit noted that any such weighing, under the law of the Ninth Circuit which governed, is only if the work is transformative, which was not the case here. In particular, the court held that “Google has conceded both that it could have written its own [versions of the code] and that the purpose of the copying was to make Android attractive to programmers.” Indeed, the court found that “there is no inherent right to copy in order to capitalize on the popularity of the copyrighted work.” The jury, however, considered that the amount taken, i.e., the 37 packages having 11,500 lines of code, were insubstantial compared to the entire body of JAVA code. The irrelevancy of this argument is staggering, e.g., under this logic you take an entire chapter or a page in a book, but the court held that the jury’s view, albeit unreasonable, makes this factor neutral.
Finally, the fourth fair use factor involves the effect on the market. As noted, Google made enormous profits with Android, which put in jeopardy Oracle’s licensing. With the variety of copyright rights imbued in a work, a copyright owner has the total or exclusive right of control of the use of the work. The facts of this case, however, were complicated and confusing. At the time of the taking, Oracle was not a manufacturer and it was unclear what value their licenses would have at the time of the taking. The court held, however, that as the copyright owner, Oracle had the right at least for the future licenses, e.g., with Google that used the technology. Indeed, the court held that the “unrestricted and widespread conduct of the sort engaged in by” Google undermined Oracle’s market opportunities, which deleteriously affected the use of the original content and any derivative works. This fourth factor, therefore, weighed heavily against fair use.
In the Federal Circuit’s final analysis of the four factors, they again noted that Google could have written their own code or properly licensed with Oracle, but instead chose to copy. “There 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.” Accordingly, the Federal Circuit held that Google’s use of the Oracle code was not a fair use.
As noted, this case is on its way to the Supreme Court with a possible trip first to an en banchearing at the Federal Circuit. Although Federal Circuit Judge O’Malley’s (and Plager’s and Taranto’s) arguments and analysis comport to the principles of copyright law, the Supreme Court may have a different balancing and take on fair use. Google is a huge corporation and their counsel quite clever. With the stakes so huge, the Court will hear this matter in a year or two, even though the total non-transformative nature of the copying by a competitor is quite telling. The numerous subtleties employed at trial likely confused the jury to no end, making them susceptible to persuasion and enamored of the trendy “transformative” arguments, even if there was no such thing. For example, an argument was made that it was transformative that only the 11,500 lines of JAVA code were chosen in the taking, the creativeness of this selection constituting a sufficient transformation under the fair use doctrine.
Clearly, the verbatim copying is a no-no, particularly if you are in a competitive area and the products are similar. Had this not been Google, this case would have been decided long ago. For those in the real world, great care is needed when appropriating another’s paradigm. Here, for whatever reason, Google gave up on writing their own 11,500 lines of code – how many programmers do they have? Clean room or other procedures are readily available to work around this problem, rather than giving up and condemning software as not protectable under copyright.
Do not take the easy way out, copy the code and hope that you will not be discovered. The copyright penalties are too great. For example, the willfulness of Google’s copying has been proven. Under copyright damages law, this could translate into $150,000 per instance of copying. Even if the normal statutory damages range of $750 to $30,000 per instance, how many Android phones are out there? Tablets? Do the math and you can see that many, many billions of dollars are at stake.
I hope that this will turn out to be a cautionary tale of what not to do. As a former computer scientist, I was aghast that the level of copying here was deemed fair use by the jury. Hopefully, the Supreme Court will see the facts similarly as the Federal Circuit and put to rest the application of the transformative doctrine in direct competing contexts.