Why a National Emergency Library Would Have Been Unnecessary
Last week, in response to the COVID-19 pandemic, the Internet Archive announced the National Emergency Library (“NEL”), which expanded digital access to the books in its collection. The New Yorker welcomed it as “a gift to readers everywhere.” Predictably, the Authors Guild, the Copyright Alliance, and the Association of American Publishers condemned the move as infringing copyright. Overlooked in this controversy is that had the 2008 attempted settlement of the litigation over the Google Library Project been approved by the court, the NEL would likely have been unnecessary.
The Internet Archive has digitized approximately 1.4 million books in its collection. It typically lends these books on an “own to loan” basis; if it owns one physical copy of a title, it will lend only one digital copy of that title at a time. After two weeks, the copy automatically disappears off of the user’s device and the title is available for another user to borrow.
To reduce its potential copyright liability, Internet Archive has generally excluded titles published within the last five years from its digital lending program. Moreover, it has always permitted authors or publishers to opt-opt by requesting that a title be removed from the digital library.
The COVID-19 pandemic has forced almost all libraries across the country to shut down, for the first time in the nation’s history. In response to the sudden need of scholars, students, and members of the public for access to books they previously would have borrowed from libraries, the Internet Archive decided to temporarily suspend its “own to loan” policy. So long as the Coronavirus emergency exists, the Internet Archive’s NEL will not restrict the number of users who can borrow a title to the number of copies it owns. However, the Internet Archive is continuing to exclude titles published within the last five years, and to permit authors and publishers to opt out of the NEL.
Rights holder organizations such as the Authors Guild have always objected to the Internet Archive’s digital lending program, contending that it exceeded the scope of fair use. The level of opposition has increased with the expansion of digital lending under the NEL. While the rights holder organizations recognize the severity of the pandemic, and the need for expanded access to content, they argue this expanded access shouldn’t be at the expense of the copyright owners, and that the Internet Archive should pay for it.
In fact, for many of the titles, there is no mechanism for the Internet Archive, or any other library, to license emergency access to the books in its collection. There are thousands of publishers, and millions of authors and titles. And for many of the 20th-century titles still under copyright, it is unclear who has the digitization rights: the authors or the publishers. Thus, there is no practical way for a library to clear the digitization rights for the books in its collection.
This was the challenge that Google faced when it started its Library Project in 2004. (See here for a detailed discussion of the Google Library Project.) It wanted to scan the full text of the world’s books into its search database, but there was no efficient means for it to clear the digitization rights in so many books. Accordingly, it designed its project to fit within the confines of fair use. It scanned books borrowed from its partner libraries, such as the University of Michigan library, into its search database. Then, in response to a search query, Google would display just “snippets” of text—a handful of lines—for the first three times a search term appeared in any given book.
In 2005, Google was sued separately by the Authors Guild and a group of five publishers. In 2008, the parties announced a sweeping and incredibly complex settlement agreement. The parties reached an amended settlement agreement in 2009. Though ultimately unsuccessful, the attempt at settling the case leveraged the fact that the rights holder plaintiffs’ claims had been presented as a federal class action. The settlement would have functioned as follows.
The amended agreement would have established a Book Rights Registry (“BRR”) that would have had the authority to collectively manage the U.S. copyrights of all books published in the United States, the U.K., Canada, or Australia before 2009. The BRR would have licensed to Google the right to offer three services relating to these books. First, under the Preview service, Google would have been able to display to a user up to 20% of an out-of-print book’s text in response to a search query, but none of the text of an in-print book. Second, Google would have been able to sell a consumer perpetual digital access to a title. Third, and most relevant here, Google could have sold institutional subscriptions to public libraries and higher education institutions. The institutional subscription database would have included the full text of all of the out-of-print books Google had digitized.
Google would have set the prices it charged for these services, within parameters established by the settlement. Google would have provided 63% of the revenue it collected to the BRR, which in turn would have distributed it to authors and publishers.
The institutional subscription would have provided a framework for libraries to enable their users to access books during an emergency such as the COVID-19 pandemic. The institutional subscription database would have contained far more titles than are in the NEL. Further, the settlement would have allowed the BRR to authorize Google to permit remote access to public library users (remote access was already permitted to higher education institutional subscribers). Significantly, the settlement would have provided the institutions, Google, and the BRR with the flexibility to make other adjustments to respond to the pandemic, such as the pricing of the subscription and the titles included in the institutional subscription database.
Unfortunately, the judge presiding over the litigation rejected the settlement in 2011. The cases were fashioned as class actions so that the settlement would apply to all class members, i.e., the copyright owners of all the books published in the U.S., U.K., Canada, and Australia prior to 2009. Class action settlements must be approved by the presiding judge so as to protect the interests of all class members. The settlement here was controversial for a different reason; some argued that it was an abuse of the class action procedure, because the settlement was broader than the activities that precipitated the litigation. Importantly, both the U.S. Copyright Office and the U.S. Department of Justice opposed the settlement on this basis.
After the rejection of the class action settlement, Google quickly reached a settlement with the five initial publisher-plaintiffs concerning the scan and snippet display of the Library Project, under which Google continued to offer this service. The Authors Guild, however, pursued litigation, resulting in two successive losses in federal courts, and yielding the Second Circuit’s decision that the Library Project was a fair use.
The Second Circuit’s decisions in the Google case and the related HathiTrust cases against Google’s library partners have provided helpful precedents concerning fair use and artificial intelligence. However, approval of the settlement would have yielded something far more ambitious. Users in multiple jurisdictions would have had far more digital access to 20th-century books, whose authors and publishers would have received additional revenue streams. Moreover, the institutional subscription available under the settlement would have rendered the NEL unnecessary.