The Future of Fair Use After Google Books
This weekend I participated in a panel about Judge Chin’s decision in the Google Books case at the Copyright Society’s mid-winter meeting. My debating partner was Jon Baumgarten, former General Counsel of the Copyright Office and partner at Proskauer Rose, where he represented publishers, among other rights holders. Jay Dougherty, a professor at Loyola Law School, acted as the moderator/referee. The discussion revealed profound disagreement on whether fair use jurisprudence is headed in the right direction.
After Professor Dougherty provided a quick overview of the case (eight years in five minutes!), I made four arguments why Judge Chin correctly concluded that fair use permitted Google’s scanning of 20 million books into its search index.
First, in support of his finding that Google Books served the public interest, Judge Chin cited five times an amicus brief I helped write for the Library Copyright Alliance.
Second, the decision is consistent with “transformative use” decisions in the Second Circuit (Bill Graham Archives v. Dorling Kindersley, Authors Guild v. HathiTrust) as well as other circuits (A.V. v. iParadigms LLC, Kelly v. Arriba Soft Corp., Perfect 10 v. Amazon.com).
Third, to the extent that Jon Baumgarten might disagree with the transformative use jurisprudence generally, the Google Books decision fell within a subset of those decisions that involved digital technology. In these cases, although there was a large amount of copying in toto, that copying occurred “behind the curtain” in the course of creating a new product or service, and the end user generally could not see entire works. Thus, the use did not supersede the market for the works. Here, for example, Google displayed only three short “snippets” in response to each search query.
Fourth, even if one had doubts about these technological fair use cases, the Google Books case presented a unique set of circumstances that distinguished it from most other cases. In particular, because the Authors Guild did not seek preliminary relief, and did not insist that Google stop scanning during the lengthy negotiations and legal proceedings concerning the settlement, Judge Chin was confronted with a highly useful database containing 20 million books. Fashioning an appropriate remedy would be exceedingly difficult. Enjoining the use of the database until Congress develops a mass digitization exception, as the Authors Guild proposed in the HathiTrust case, was a completely unrealistic remedy. Thus, finding fair use may have been the best way forward given the facts on the ground.
Jon Baumgarten then explained why he thought Judge Chin decided the case incorrectly. Jon agreed with me that there was precedent supporting the decision, but argued that Judge Chin ignored other, more compelling, precedent. Jon’s basic point was that Judge Chin compounded the errors that are the heart of many of the transformative use cases: rather than identify whether the expression is transformed, as Judge Leval originally intended, these cases simply ask if the expression is being used for a new, socially beneficial purpose. Jon is concerned that it will be too easy for clever lawyers to highlight a socially beneficial purpose, as my amicus brief did in this case. The emphasis on socially beneficial purpose changes the focus from the infringer’s actions to uses made by third parties. It limits the ability of rights-holders to receive financial benefits from the use of their works. Analog works will become fodder for technology companies.
Jon added that although fair use is extremely important to the functioning of the copyright system, it should involve incidental activity, and not be conducted at the enterprise level. Under Judge Chin’s analysis, Google is “too big to infringe.”
Finally, Jon noted that Google’s “Partner Program,” under which it entered into licensing arrangements with publishers, indicated that there was a functioning licensing market, tilting the fourth factor against Google.
In my limited rebuttal time, I responded that while Google reached licensing agreements with some publishers with respect to some books, the copyright ownership of many of the books scanned by Google was unclear. The copyright might belong to the publisher, the author, or the author’s heirs. The transaction costs involved in determining the ownership of these books’ copyrights made licensing infeasible.
I added that reliance on fair use in this case, as well as in many of the other transformative use cases, was a function of the conflicts resulting from the interaction of a changing copyright system and evolving digital technology. On the one hand, the lengthening of copyright term and the abandonment of formalities as a condition for protection means that more works are covered by copyright than ever before – including many of the works at issue in this case. At the same time, the nature of digital technology means that users are making dramatically more copies in the course of their daily activities – turning on their computer, viewing websites with browsers, responding to and forwarding emails. During the course of the day, a user may make hundreds, if not thousands, of copies. Fair use is the theory that permits these activities.
Professor Dougherty asked the first question: was the Google Books decision consistent with the Berne three-step test? Jon Baumgarten said that it may well not be; how could the copying of millions of books be a “special case?” I replied that the three-step test applies to statutory exceptions a national legislature may adopt, so the issue would be whether fair use is compliant with the three-step test, not whether this application of fair use is three-step compliant. I noted that ever since joining the Berne Convention, the U.S. government has taken the position that fair use is three-step compliant, and as the world’s only superpower, our position on this issue is dispositive. Further, the second and third steps (“do not conflict with a normal exploitation of the work and do no unreasonably prejudice the legitimate interests of the author”) sounded like the first and fourth fair use factors (“the purpose and character of the use,” “the effect of the use upon the potential market for or value of the copyrighted work”).
A member of the audience asked whether the Google Books and HathiTrust decisions rendered the updating of section 108, as recommended by the Copyright Office, superfluous. I responded that while the exceptions in section 108 were extremely important to the operation of libraries, they probably were a bit out of date, as was the rest of the Copyright Act. Fortunately, fair use can provide some flexibility to specific exception. I mentioned my article published in the Journal of the Copyright Society where I argued that in situations where a defendant engaged in the sort of activity permitted by a specific exception, but ultimately did not qualify for a narrow technical reason, the court should consider the defendant’s substantial compliance with the exception when applying the first fair use factor – the purpose and character of the use. I added that in a perfect world, Congress should update section 108, and the rest of the Copyright Act. But we live in an imperfect world, and given the availability of fair use, it made no sense for libraries and publishers to spend five years negotiating an update to section 108 that would be obsolete when completed.
Jon Baumgarten opined that my Copyright Society article was very scholarly, and dead wrong.
Another member of the audience made an impassioned statement that by pushing the fair use envelope, Google was in effect deciding unilaterally what the copyright law was, thereby forcing copyright owners to abandon their rights. Some audience members applauded at the conclusion of this statement.
I replied that it wasn’t only Google that was relying on fair use. Some recent prevailing fair use defendants include Bloomberg, Thomson Reuters, and Reed Elsevier. I pointed out that Thomson Reuters and Reed Elsevier without authorization copied over a million briefs and other pleadings, which they incorporated into a commercial database product. When sued by a law firm, these publishers successfully argued that they had engaged in a transformative use.
Moreover, in many cases, plaintiffs are able to convince courts to reject assertions of transformative use. In the Salinger case, the Second Circuit found that a sequel to Catcher in the Rye was an infringing derivative work. Likewise, a district court found that a Harry Potter lexicon copied more expression than necessary to accomplish its purpose. In other words, courts are perfectly capable of distinguishing legitimate transformative uses from infringing derivative ones. They are readily equipped to find uses that will advance the public interest without diminishing an author’s incentive to create.
After a panel of this sort, I always think of arguments I should have made, or more convincing ways to frame an argument I did make. A point I did not make is one I wrote about soon after the decision. I argued that for at least some of the Authors Guild members, this case has always been about the principle that no one should use their works without their permission. While they may agree with fair use in the abstract, they oppose it as applied to their works. The fact that the use is socially beneficial and does not harm them economically is irrelevant. They believe that creators should have complete control over copies. The positions Jon Baumgarten expressed this past weekend are consistent with this perspective.
This belief in complete control is based more on the Continental “author’s rights” (droit d’auteur) tradition than on the Anglo-American utilitarian tradition. In the author’s rights approach, copyright springs not from statutes but from natural law. By contrast, in the Anglo-American system, copyright is not a response to natural law, but rather is a matter of legislative choice directed at incentivizing the creation of works for the benefit of society. The Anglo-American utilitarian approach in theory provides only as much protection as is necessary to encourage creative activity, while the author’s rights approach provides more robust protections of both economics and moral rights such as the rights of integrity and attribution.
However, in response to lobbying by right-holders, Congress has enacted certain features of author’s rights systems – for example, the ever-increasing copyright term, referenced above. The first U.S. copyright act provided a term of 14 years, renewable for another 14 years, for a maximum total of 28 years. Now, the copyright term matches the European Union’s term of life of the author plus 70 years.
The complete control over copyrighted works sought by the Authors Guild in this case is inconsistent with the public interest purpose of our copyright system. Fortunately, Judge Chin recognized that the objective of copyright is not to enrich rights-holders, but “to advance the progress of the arts and sciences.”
Jonathan Band is a DC-based attorney whose clients include Internet companies, providers of information technology, universities, library associations, and CCIA. He previously guest-posted on DisCo about the impact of software infringement on manufacturing competitiveness.