Judge Rules against Internet Archive; Will AI Be Next?

A US District Court judge in New York has issued a ruling granting a motion for summary judgement filed by book publishers against the Internet Archive (IA), alleging copyright infringement.

According to the opinion, the four publishers (Hachette Book Group, Inc., HarperCollins Publishers LLC, John Wiley & Sons., Inc., and Penguin Random House LLC) allege that the non-profit defendant, “an organization whose professed mission is to provide universal access to all knowledge,” infringed the copyrights in 127 books published by the plaintiffs.

The defendant scanned print copies of the works and lent the digital copies to borrowers via its website without the publishers’ permission.

The defendant argued that this was permitted by the “fair use” exception to copyright law under 17 U.S.C. § 107, which states:

the fair use of a copyrighted work, including such use by reproduction in copies or phonorecords or by any other means specified by that section, for purposes such as criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research, is not an infringement of copyright. In determining whether the use made of a work in any particular case is a fair use the factors to be considered shall include—

(1) the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;

(2) the nature of the copyrighted work;

(3) the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and

(4) the effect of the use upon the potential market for or value of the copyrighted work.

The judge noted that

One of IA’s first projects was to document the history of the Internet by archiving every public webpage on the World Wide Web through IA’s “Wayback Machine.” … IA also works with libraries, museums, universities, and the public to preserve and offer free online access to texts, audio, moving images, software, and other cultural artifacts.

The dispute concerns the way libraries lend digital ebooks.

Libraries usually buy print books from publishers or wholesalers. However, ebooks are typically not bought but licensed to libraries from publishers through distributors called “aggregators.”

Publishers require aggregators to ensure that libraries lend ebooks only to the library’s members. Publishers also require aggregators to use “digital rights management” (“DRM”) software and other security measures to prevent unauthorized copying or distribution of ebooks.

IA’s website includes millions of public domain ebooks that users can download for free and read without restrictions. However, the site also includes 3.6 million books protected by copyright, including 33,000 of the plaintiffs’ titles.

IA doesn’t make its ebook copies available for mass download. Instead, it lends only limited numbers of these works at a time through “Controlled Digital Lending,” or “CDL.”

This means that an entity that owns a physical book can scan that book and “circulate [the] digitized title in place of [the] physical one in a controlled manner,” on a one-to-one “owned to loaned ratio”:

For example, if a library owns three copies of a title and digitizes one copy, it may use CDL to circulate one digital copy and two print, or three digital copies, or two digital copies and one print; in all cases, it could only circulate the same number of copies that it owned before digitization.

Around 2018, IA began expanding its lending capacity through its “Open Libraries” project which let libraries pool their physical collections using IA.

As the court noted,

Anyone can become a patron of IA, and borrow up to ten ebooks at a time for up to fourteen days each, by submitting a valid email address…. IA never charges patrons fees for any service, including borrowing books.

However, when in March 2020 the Covid-19 pandemic closed libraries nationwide, by IA’s estimation that took 650 million print books out of circulation.

IA then launched what it called the National Emergency Library (“NEL”), lifting the technical controls enforcing its one-to-one owned-to-loaned ratio and allowing up to ten thousand patrons at a time to borrow each ebook on its website.

IA ended the NEL on June 16, 2020, soon after the publisher’s lawsuit was filed, and went back to its prior lending model.

Since then, IA’s user base has increased from about 2.6 million to 6 million, with about 70,000 ebooks borrowed daily.

The court reviewed the four fair use factors listed above and concluded that IA’s use was outside the scope of fair use, noting: “There is nothing transformative about IA’s copying and unauthorized lending of the Works in Suit.”

The court distinguished the Google Books case, in which the defendant

scanned whole copies of millions of books, including those protected by valid copyrights, to create a database on which the general public could search for particular terms across the scanned works.

There, said the court,

The creation of this “full-text searchable database [was] a quintessentially transformative use,” … because “the result of a word search is different in purpose, character, expression, meaning, and message from the page (and the book) from which it is drawn.”… Importantly, the database did not “allow users to view any portion of the books they [were] searching” and therefore, unlike IA’s Website, “d[id] not add into circulation any new, human-readable copies of any books” or “merely repackage or republish the originals.”

IA argued that its library was “wholly noncommercial” because it’s a non-profit organization. However, noted the court, “The absence of a dollars and cents profit does not inevitably lead to a finding of fair use.”

Said the court,

Although it does not make a monetary profit, IA still gains “an advantage or benefit from its distribution and use of” the Works in Suit “without having to account to the copyright holder[s],”…. The commercial-noncommercial distinction therefore favors the Publishers.

IA also argued that its lending was allowed under the first sale doctrine, which provides that a

rights holder’s control over the distribution of any particular copy or phonorecord that was lawfully made effectively terminates when that copy or phonorecord is distributed to its first recipient.” … Thus, “the lawful purchaser of a copy of a book is free to resell, lend, give, or otherwise transfer that copy without violating the copyright holder’s exclusive right of distribution,” and “[t]he copy so resold or re-transferred may be re-transferred again and again without violating the exclusive distribution right.”

The court rejected this argument, noting “IA points to no case authorizing the first recipient of a book to reproduce the entire book without permission.”

The IA decision may also influence pending fair use cases involving AI, in which IP owners have argued that their images and other content are being copied without permission to train AI engines to generate “new” derivative content.

We recently blogged about AI IP issues here and here.