Copyright Office Questions Legality of Internet Archive’s National Emergency Library

On March 24, the Internet Archive (Archive) unveiled what it called the “National Emergency Library” (NEL) in order to “address our unprecedented global and immediate need for access to reading and research materials.” The announcement specified that Archive would suspend the waitlist for 1.4 million books in its unlicensed “lending library” until at least June 30, thus allowing an unlimited number of people to download electronic copies of the same book at the same time. Archive had previously employed a controlled digital lending (CDL) model where the number of downloads was tied to the number of physical copies Archive or its partners possessed. With the waitlist suspended, Archive temporarily abandoned the CDL model that it had relied on since 2011.

Archive’s release of 1.4 million copyrighted works without a license certainly caught people’s attention. The Authors Guild quickly condemned the move, claiming that Archive “has no rights whatsoever to these books, much less to give them away indiscriminately without consent of the publisher or author.” Maria Pallante, President and CEO of the Association of American Publishers, likewise denounced Archive’s announcement: “We are stunned by the Internet Archive’s aggressive, unlawful, and opportunistic attack on the rights of authors and publishers in the midst of the novel coronavirus pandemic.” Archive then responded in a blog post explaining that it had suspended its waitlist due to the “tremendous and historic outage” in the nation’s libraries caused by the pandemic and arguing that fair use is the “legal doctrine underlying” its CDL model “during normal times.”

A couple of weeks later, on April 16, Senator Tom Udall (D-NM) sent a letter to Acting Register of Copyrights Maria Strong asking the Office “to examine the National Emergency Library that has been organized by the Internet Archive which is operating without typical library licenses and is causing authors in New Mexico concern about the integrity of their copyrights.” In particular, Sen. Udall asked the Office to “include a legal analysis of the Internet Archive’s National Emergency Library” under Section 107 and to “recommend any corrective action that you deem necessary to comply with copyright law and protect authors.”

On May 15, Acting Register Strong submitted a detailed response to Sen. Udall, noting that “it is not the Office’s general practice to provide legal advice about specific factual scenarios” and that the “Office is particularly cautious about weighing in on circumstances or disputes between private parties.” Nevertheless, the Office provided a general analysis of how copyright law applies to libraries and then looked at how that analysis applies to “the Internet Archive’s recent activities.” Notably, the Office ultimately concluded that it “would have been beneficial for the Internet Archive to engage with writers and publishers prior to launching the National Emergency Library to discuss the contemplated parameters for the project and determine their willingness to participate.”

On June 1, a couple of weeks after the Office submitted its response to Sen. Udall, four major publishers, including Hachette, HarperCollins, Penguin Random House, and Wiley, filed suit against Archive for copyright infringement in the Southern District of New York. The complaint, which includes 127 works-in-suit, alleges that Archive’s CDL and NEL models infringe on their works, both directly and indirectly. This blog post does not address that dispute, though the publishers do raise many of the same issues in their complaint that the Office raised in its response to Sen. Udall. This blog post merely summarizes the Office’s reasoning on the fair use analysis of Archive’s National Emergency Library. It is worth noting that, even though Archive announced on June 10 that it was shutting down its NEL, the legality of the NEL is still a live issue in the publishers’ lawsuit.

Fair Use Under Section 107

Looking at the first fair use factor under Section 107 generally, the Copyright Office notes that while the “goals of promoting scholarship and education are explicitly identified in the statute as favored purposes,” it “is generally understood that many uses of copyrighted works by schools and universities must be licensed.” Citing Oracle v. Google and Authors Guild v. HathiTrust, the Office points out that “reproducing the text of physical books in digital format is not transformative unless the change in format results in new uses for the work.” Moreover, it explains that using educational materials for educational purposes “would not serve a different purpose than the original.”

Turning to the NEL specifically, the Office takes issue with Archive’s claim that “the vast majority” of the books it makes available “do not have a commercially available ebook” that would be publicly available given that libraries are closed. On the contrary, the Office states that “Archive does not appear to have verified if any of the works in its collection were available to the public in digital formats prior to including those books in its collection or removing its waiting lists” and that the NEL “includes many books for which ebooks are available commercially” at local libraries. Thus, the argument that Archive was making available works that were otherwise unavailable “does not apply to any books that were available in digital formats at the time of the copying.”

The Office notes that the NEL is available to the public for free such that Archive’s use is noncommercial. However, it emphasizes that Archive’s stated purpose of promoting scholarship and education “alone does not establish fair use.” Indeed, the Office points out that “at least some” of the 1.4 million works, such as “Stephen King thrillers and joke books,” are “likely to be accessed for entertainment rather than educational purposes.” Even for the educational books that are not available in digital formats, the Office explains that the noncommercial purpose must be weighed against “the non-transformative nature of the use.” Given that educational works are “originally intended to educate,” Archive’s use of these works “is not transformative.” And given that Archive does not provide “search functionality,” it does not fit within the “digitization cases” that were “deemed transformative,” such as Google Books and HathiTrust.

On the second factor, the Office cites Campbell v. Acuff-Rose for the proposition that “some works are closer to the core of intended copyright protection than others” and Harper & Row v. Nation for the point that the “law generally recognizes a greater need to disseminate factual works than works of fiction or fantasy.” The Office also notes that there may be “more justification” for reproducing previously published works that are “currently unavailable in the marketplace.” Nevertheless, it explains that the existence of organizations to provide copies of such works is relevant to this factor, and it mentions that the case law on a “work’s print status under the second factor is mixed.”

Applying this to the NEL, the Office clarifies that the analysis of the second factor is necessarily “fact-specific” and that each work or category of works “would need to be evaluated independently.” For example, “creative works” would be analyzed differently than “factual or informational works.” And while the unavailability of certain works might favor fair use “in some circumstances,” the Office notes that this does not appear to have been Archive’s focus. Archive instead focused on whether the books were available in digital form, making no “mention of the works’ overall availability.”

The key to the third factor, the Office explains, is “whether the secondary use employs more of the copyrighted work than is necessary, and whether the copying was excessive in relation to any valid purposes asserted under the first factor.” Moreover, while “copying an entire work often weighs against a finding of fair use,” the factor “would not weigh against a finding of fair use” if “it were necessary to copy the entire copyrighted work to achieve the purpose of the secondary use.”

The Office notes that the CDL White Paper, upon which Archive relies, “argues that it is necessary to copy the entire book to achieve the purpose of providing digital access to the work, such that the copying is not excessive in relation to the library’s purpose.” The CDL White Paper also argues that “the library prevents users from making additional copies of or further distributing the book and limits the duration for which a user can access a book.” In response, the Office points out that the courts in Google Books and HathiTrust “emphasized that the defendants had not made the full text of the copied works visible to the public” and how the Office itself (see here, here, and here) has “consistently expressed doubt that providing digital access to complete works can be considered a fair use.”

Quoting Harper & Row, the Office notes that the fourth factor, which is “undoubtedly the single most important element of fair use,” turns on “whether widespread conduct similar to the conduct of the alleged infringer ‘would adversely affect the potential market for the copyrighted work.’” The Office explains that, under Google Books and HathiTrust, the third and fourth factors are linked: “the risk that the digitized version will serve as a market substitute for the original work increases as the amount of the work that is made accessible to the public increases.” In both of those cases, the copying of entire works was permissible because it enabled transformative search functionality without serving as a market substitute by making the entire works available.

Turning to the NEL, the Office acknowledges a “disagreement among stakeholders over whether the analysis of market harm under the fourth fair use factor should consider the Internet Archive’s activities as roughly analogous to physical lending by libraries, or whether the markets for physical lending and ebook licensing to libraries are distinct.” The Office points out that no court has embraced the former approach and that “the Second Circuit squarely rejected it” in Capitol Records v. ReDigi. Moreover, the Office cites its own report noting that there are “significant differences” between lending physical copies and digital ones. And it points out that the NEL “lacks the controls cited by the CDL White Paper as necessary to mitigating market harm” since it allows “an unlimited number of users to borrow any given title simultaneously.”

On the latter approach, which holds that the market for ebook licensing is distinct from that of physical lending, the Office states that there is already “an established market” where “publishers and authors license their works to libraries for the purpose of digitally ‘lending’ them to patrons.” The fourth factor analysis here “might focus on whether the creation and distribution of digital versions of these works would affect this market, and also how, if such conduct became widespread, it would affect this market.” If digital versions of some works were not available in the marketplace, the Office concludes, “this factor might favor fair use for some, but not necessarily all, of the works contained in the National Emergency Library.”

The Office also examines how “exigent circumstances” related to the pandemic may factor into a fair use analysis. The Office notes that there “is undoubtedly a strong public interest in ensuring continued access to educational materials in this unprecedented time, which could weigh in favor of fair use.” However, while Archive’s goal “may be laudable, so is respect for copyright.” The Office states that it “would be imprudent to excuse widespread copying due to a national emergency without considering the possible repercussions on copyright law and copyright owners” since there “is also a strong public interest in ensuring that authors are able to financially survive the coronavirus crisis to be able to continue to produce creative works.” And it concludes by noting that a “court would almost certainly also take into account” Archive’s effect “on writers and publishers.”

Conclusion

The Copyright Office ultimately suggests that Archive may wish to “explore opportunities for collaboration with writers and publishers” such as by “allowing them to opt into making digital versions of their works publicly available.” While the Office never explicitly says that any particular work is being infringed, its analysis does indicate that, in its opinion, Archive cannot claim that it is engaging in fair use for all of the works made available with its National Emergency Library. Furthermore, the Office explains why other statutory limitations, such as the first sale doctrine under Section 109, would not apply to Archive’s activities. All in all, it seems clear that the Office is quite skeptical of the NEL’s legality under the fair use doctrine. Though, it remains to be seen whether the Southern District of New York will agree with the Office’s analysis.