Copying practices in the academic world were again thrown into legal disarray when a federal appellate court reversed a fair use finding in favor of Georgia State University (GSU) in its long-standing copyright dispute with several academic publishers. The trial court had found the practices, which involved scanning portions of books into electronic course reserves and other systems for classroom use, to be a fair use of the copyrighted works. The appellate court’s 129-page opinion determined that the trial court had not properly applied fair use law, but stopped short of declaring GSU’s practices to be illegal infringement. The appellate court sent the case back to the trial court for a reconsideration of GSU’s copying.
The Course Pack Cases
The practice of selecting excerpted readings for use in classrooms has a long history. In the analog days, teachers would select readings for classroom use and arrange the creation of a “course pack”—a copied and printed collection of readings that would be compiled and distributed to students. When the Copyright Act of 1976 was being formulated, the U.S. House of Representatives developed a series of agreements known as the Classroom Guidelines. These guidelines established certain minimum “safe harbors” for copying for classroom use, including rules that specified the copying of sources such as a single chapter from a book, articles of 2,500 words or fewer, or book sections or chapters of up to 1,000 words or 10% of the work.
The guidelines did not have the force of law and were tested in two court decisions (Basic Books, Inc. v. Kinko’s Graphics Corp. and Princeton University Press v. Michigan Document Services, Inc.) known as the course pack cases. Both involved the creation of course packs, which generally included chapters and readings from several books. Notably, while the instructors identified the readings to be used, the compiling and printing of the course packs were done by commercial copy shops near the campuses. In both cases, courts found that fair use did not apply because the copying was done by the commercial copy shops. They did not address whether fair use would have applied if the copying had been done by the nonprofit universities.
After those decisions, many universities created policies that paid royalties to copyright owners for the material selected and used in the course packs, even when the course pack copying was done by the university. As course packs transitioned from copied and printed analog packets to scanned and uploaded images on e-reserve and course websites, it became less common for the royalties to be paid. According to testimony at the trial, GSU faculty considered that “digital copies of excerpts consisting of up to 20 percent of a work” could be posted without obtaining a license. The limited availability of licensing systems was also asserted to be a contributing factor in GSU’s practices. The trial court determined that while licenses or a permission structure were frequently available for print copies and course packs, licenses were less readily available for digital excerpts.
In reaction to this trend, a coalition of three academic publishers (Cambridge University Press, Oxford University Press, and SAGE Publications) sued GSU in 2008 for copyright infringement. After extensive proceedings and a lengthy trial, the court ruled that most of the copying that GSU was engaging in was a permitted fair use. The Fair Use doctrine, found at Title 17, Section 107 of the U.S. Code, allows for the free use of copyrighted works for certain purposes, including teaching, research, and scholarship. The reasoning behind the doctrine is to allow for new knowledge to be built on a foundation of existing copyrighted knowledge.
Applying the Fair Use Doctrine
The Fair Use doctrine famously outlines four factors to be considered in determining whether a use is fair. These are the nature of the new use, the nature of the original work, the amount of the work being used, and the impact of the new use on the market for the original work. In ruling that GSU’s copying was a fair use (in most cases), the court found that the first factor “strongly favored” fair use because the use was for “nonprofit educational purposes.” Because virtually all of the works being scanned were “informational in nature,” the second factor also favored fair use. Considering the amount of the work being used, the trial court set a standard of 10% or one chapter as a fair use safe harbor. In looking at the fourth factor, the trial court determined whether “digital permissions were readily available for excerpts” of books. If they were not, then that factor favored fair use. The district court then added up the factors for each of the works and found that three or more established fair use.
Interestingly, the appellate court agreed with a number of the specific findings of the trial court, while determining that, nonetheless, the trial court’s ultimate conclusions were incorrect. While agreeing that the nonprofit educational purpose of the copying supported fair use, the appellate court expressed concern that the use was not “transformative,” in that it achieved the same educational purpose as the original work. Because of this, the first factor carries less weight in the overall fair use decision. Similarly, while the trial court essentially concluded that the works being copied were “informational,” the appellate court said that it wasn’t a hard-and-fast rule. Informational works can incorporate greater or lesser levels of the “authors’ perspectives and opinions” that can impact the fair use evaluation. The appellate court rejected the trial court’s blanket finding in favor of fair use, indicating that the works need to be individually evaluated.
The appellate court also rejected the trial court’s safe harbor rule of 10%, saying that a “work-by-work” analysis was required. In conducting that analysis, the court considered the pedagogical purpose of the use, the entire length of the book—including the preface, table of contents, index, and similar material—and that the analysis was not bound by the Classroom Guidelines. Finally, the appeals court agreed that the trial court had correctly considered the impact of licensing availability and license revenue in determining the fourth factor’s market impact. However, because GSU’s use was not transformative, not only was the first factor given less weight, but this fourth factor was given greater weight.
What the GSU Decision Means
In summary, the appeals court agreed with much of the trial court’s fair use analysis, but not with how it applied that analysis. The trial court erred “by treating the four factors mechanistically”; “a holistic analysis which carefully balanced the four factors” was required.
Sending the case back to the trial court does not necessarily mean that the outcome will change. The trial court could come to the same conclusions based on these new instructions from the appeals court on how to apply fair use’s four factors. Indeed, the early commentaries on the case suggest that while this was a positive decision from the perspective of the publishers, it may not have been the decision they wanted. By emphasizing a work-by-work analysis and rejecting safe harbor standards, the court also rejected the publishers’ position that GSU’s practices as a whole were infringing on copyright. This may make it more difficult for other publishers to challenge institutional practices as a whole.
Or perhaps not. The decision could also serve as an instruction guide for how to more effectively bring a copyright infringement claim to court. Also, one of the appellate court’s judges drafted a separate opinion (beginning on page 113 of the decision) indicating that in his view, GSU’s practices were absolutely not a fair use. This suggests that the issue of copying in the academic environment remains far from settled.