Today, the Supreme Court declined to consider the much publicized Authors Guild v Google fair use case (aka the Google Books case). In refusing the case, the Supreme Court not only let stand a Second Circuit decision that dramatically expands the boundaries of the fair use doctrine’s transformative use test, but more importantly it missed a golden opportunity to provide crucial guidance to copyright owners and users, the public and the courts as to how the fair use doctrine should apply to a whole new type of uses – functional uses – made possible by new technologies and for which the existing transformative use test in inadequate and inappropriate.
In the Google Books case the Second Circuit held that Google did not need the consent of the copyright holders of 20 million books to digitize those books in order to create a publicly searchable database containing those and other books. The court’s holding was based on a dubious finding that Google’s mass digitizing effort was a fair use because the Google Books project conveys “information” about the works to users and therefore transforms the books.
The Problem with 21st Century Fair Use Analysis and the Transformative Use Test
The problem with the outcome of the Google Books case is less about the court’s final decision and more about how it reached that decision. There are several issues with the court’s fair use reasoning in the case, but the primary problem was that the court largely relied on the transformative use test to conclude that Google’s use was a fair use and thus it did not need to get the copyright owners permission or compensate them for the making of exact copies of their books.
The transformative use test was first articulated by the Supreme Court over twenty years ago in Campbell v Acuff-Rose Music Inc. In that case the Court held that “the more transformative the new work, the less will be the significance of other factors.” Significantly, the test – as originally articulated by the Supreme Court – hinged on there being a new expressive work that was created as a result of the user’s transformative use. In cases like Google Books there is no new expressive work that results from the use. Rather, the works are being used to create a service that uses information about the works that is culled from the copying of them. Since there is no new expressive work produced as a result of the copying, the transformative use test is irrelevant. Unfortunately, many courts, including those that heard the Google Books case, have begun relying on the transformative use test as though it can serve as the sole basis for a fair use finding.
Courts have been forced into this position because the Supreme Court has only given the lower courts one tool in the toolbox and that tool is the transformative use test. It’s the typical “square peg in a round hole” problem. The crux of the problem lies with the Supreme Court’s failure to provide guidance to the lower courts in how to handle these new type of “functional use” cases which in turn has led to a critical warping of the transformative use test. The transformative use test has become so expansive and so untethered from the core principles the Supreme Court set forth in Campbell and from the goals underlying the Copyright Act that many courts have begun to treat it as though it is the most critical element of the fair use analysis, often overwhelming the other factors and leading to inequitable results.
Although Google’s use here should not be considered to be transformative use, that does not necessarily mean that Google’s use may not ultimately be considered a fair use after an evaluation of the facts and all four factors of the fair use test. It’s also worth noting that, while Google Books may be the most well known functional use case, it was not the first and won’t be the last. So if there’s a problem with the court’s reasoning in the Google Books and similar cases how do we fix that?
The answer is simple: courts need another tool in the toolbox – give them a new test – one that more effectively addresses the competing interests of the parties and the goals underlying the Copyright Act.
It’s Time for a New Functional Use Fair Use Test
What is taking place in Google Books is what is often referred to as a “functional use.” A functional use is one in which the user copies copyrighted works in order to create a product or service that performs a different function than the copyright owner’s original work. Functional uses often involve systematic, large-scale copying that generates no new expression or works. Google Books is a typical functional use case because the ultimate function of Google’s book scanning project is to create a search tool that provides information about the books, not the expression contained in the books.
The fair use doctrine is an equitable doctrine, but in functional use cases it hasn’t worked that way because the transformative use test is ill equiped to effectively balance the competing interests at stake in these cases. Fair use analysis should take into account not only the interests of owners and users but also the underlying policy objectives of the copyright law. To account for these factors in a reasonable and balanced way, it is time for the courts to begin using a functional use test. I am not the first one to suggest that courts adopt a functional use test. Indeed, many scholars, including professors and judges, have also suggested similar tests.
The functional use test would obviously not apply in every instance. Rather, it would apply to instances of actual functional uses. That is, those uses in which (1) it is necessary for the user to copy a large volume of works to accomplish the project; (2) the use results in a product or service that uses the works in a way that does not create a new expressive work; (3) the user copies entire works but only as a means to develop a product or service based on the non-expressive elements of the works; and (4) the ultimate product or service achieves a different function than the copyright owner’s works and does not compete with or serve as a substitute for those works. Any functional use test must consider all four of these factors.
In addition to considering factors relating to the type of use, a functional use test should also take into account the competing interests of both the copyright owners and the users by evaluating: the (5) ease by which the copyrighted works can be licensed from centralized licensors and (6) extent to which the user employs commercial-grade security to protect against unauthorized access to the works and, if that security system is breached, the extent to which the user takes immediate and effective action to limit the harm to copyright owners whose works were affected by the breach. These last two factors could benefit from some further explanation.
The ease by which the copyrighted works can be licensed from centralized licensors
In many instances there is an actual or potential market for the copyright owners to license their works to users to create these services. For instance, at the time the Google Book project began Microsoft had already begun negotiations with publishers for a book database. Moreover, Google itself has obtained licenses from certain publishers. The fair use doctrine should not be applied in a way that destroys or otherwise harms these markets.
On the other hand, since the new products and services created by users often require that the users engage in large-scale copying of these works, the time and costs associated with tracking down each copyright owner and licensing each work in the corpus might be so burdensome as to preclude the creation of the product or service.
Because most functional uses necessarily require a substantial number of works, maintaining the economic viability of such uses requires a mechanism to minimize transaction costs associated with identifying and licensing works, but that also ensures that copyright owners are appropriately compensated for the use. The best solution to this problem is through centralized licensors – entities that own large catalogues of works and directly license to third parties or entities that are authorized by copyright owners to license large catalogues of works to third parties (e.g., Copyright Clearance Center, Getty Images). If a user can cost-effectively license the works from a centralized licensor, then the interests of both copyright owners and users will be satisfied. Thus, where a user could license the copyright works from one or more centralized licensors, this fact should weigh against a finding of fair use as to any work that can be licensed.
But what about works that the user needs that are not included in the centralized licensor’s corpus? The fact that these works are not available to be licensed from a centralized location should weigh in favor of a finding of fair use. This approach would allow courts to better balance the interests of owners and users by allowing mixed fair use decisions. The transformative use analysis that courts presently apply is usually an all-or-nothing determination in which the court allows either all the works to be used or none of them, based simply on whether the underlying use “transforms” the works at issue in some way. It essentially presents judges with a Hobson’ choice: stop the use, which may advance significant public policy objectives, or give users a free ride. Allowing courts to take into account the ease by which the copyrighted works can be licensed from centralized licensors would give them greater ability to reach equitable results.
Balancing the interests in this way will also create incentives for small and mid-sized copyright owners to develop collective licensing models which will ultimately allow more copyright owners to get paid for use of their works while also increasing the licensing party’s corpus of works, thereby making it easier for users to get access to more works more easily.
The extent to which the user is using commercial grade security to protect against unauthorized access to the works and the extent to which the user takes immediate and effective action to limit the harm to copyright owners if that security system is breached.
It is essential that functional-use users protect the large volume of works that comprise its corpus with the same protection and care as the copyright owner would. However, under existing law, neither the fair use doctrine nor the transformative use test requires that a user protect the corpus of works used from being illegally accessed. Thus, there is no incentive for the user to employ adequate and effective security to protect against unauthorized access to or copying of the works. On the other hand, the failure to effectively protect these copyrighted works could cause tremendous harm to the copyright owners whose works get accessed and/or copied.
To balance these competing interests, users should have to take on certain security obligations in order for their uses to be considered to be a fair use under the functional use test. While the precise contours of a security apparatus would necessarily be determined on a case-by-case basis, one could imagine that a functional user might either permanently delete all copies immediately after they have been used for their intended purpose or employ a commercial-grade security system to prevent third parties from unlawfully accessing the copies. The robustness and effectiveness of the security scheme would be significant factors in determining whether the use should be allowed as a fair use.
If there is a security breach and someone illegally accesses the works, there also needs to be some onus on the user to take immediate action to prevent any harm to the copyright owners. Under normal circumstances the steps that the user would need to take would be outlined in a license agreement, but since the use was a fair use there are no such terms, and there would no incentive for the user to take such actions. To balance the interests of the parties the user must be obligated to take immediate and effective action to limit the harm to copyright owners whose works were affected by the breach. The type of action will differ depending on the facts of each use and each breach, but in all instances it should at the very least include immediately notifying the copyright owners of the breach and the steps that the user is taking to limit harm to them. The extent to which the user does or does not satisfy these requirements, should be a significant factor in determining not only whether the initial use should be allowed as a fair use but also, importantly, whether the user should be allowed to retain its fair use defense or should lose the defense, thereby opening the door for the user to be potentially liable for copyright infringement based on the initial use.
Of course, none of these factors should be considered in a vacuum. Courts should still continue to consider all the other fair use factors. The test described above is a proposal for “subfactors” that would apply specifically in cases where a court is called upon to consider functional uses – it is intended to serve as a replacement for the “transformativeness” test which courts have increasingly, and inappropriately, been applying as a one-size-fits-all approach to fair use. This test, which would be applied across both the first and fourth fair use factors, will help the courts better understand the complexities and competing interests involved in functional uses that may not be fully contemplated by the familiar fair use test.
The Supreme Court had the opportunity to fashion a new fair use test in Google Books but, by denying the case, failed to do so. There is bound to be another opportunity in another functional use case that is sure to arise in the future. Hopefully, the Supreme Court or a lower court will take that opportunity to create a functional use test modeled after the test proposed here that will allow courts to return the transformative use test to one that better reflects the Supreme Court’s original intent while also restoring balance to the fair use doctrine.