Publishers Call Internet Archive's CDL an 'Industrial Infringement Program'

The U.S. District Court for Southern New York ruled against IA in March 2023, holding it liable for copyright infringement of plaintiffs’ works, saying it worked “in active concert and participation with the Open Library of Richmond,” which scanned physical books and made them available publicly on IA's website under a “books to borrow” collection. Terms of the monetary judgment payment were confidential and conditioned on IA’s right to appeal.

In their Friday brief, publishers Hachette, HarperCollins, John Wiley & Sons and Penguin Random House said that IA claims under its CDL theory that it has the right “to appropriate” the value of physical books in digital form “without compensating creators or agreeing to standard terms." For instance, a Google search for “free ebooks” or “Toni Morrison Beloved free read” takes users to IA’s website “on the first page of the search results, above links to Publishers’ authorized library ebook platforms,” said the brief.

When the lawsuit started in June 2020, IA’s website offered 1.3 million in-copyright books; that number grew to 3.2 million two years later, including more than 33,000 of the publishers’ titles available in print and digital form, the brief said. As of February 2022, IA “effectuated roughly 25 million 'borrows’ of in-copyright ebooks per year,” it said.

On IA, users can read entire works on tablets and e-readers in “high-quality” PDF and ePUB versions, said the brief. From 2017-2020, the 127 works at issue were checked out more than 46,000 times on the IA website, it said. Extrapolating the numbers across 33,000 of the publishers' titles on the website in print and digital form, “the data suggests” that IA’s “bootleg versions” of the publishers’ ebooks were checked out “many millions of times” during three years, the brief said.

Over the past 10 years, IA has “actively enlarged the scope of its distributions,” said the brief. IA said in 2007 it would begin scanning out-of-print works to be distributed through a digital interlibrary loan system; by 2011, it “had abandoned” the out-of-print restriction “but limited itself to distributing ebooks on an ‘in-library lending model,'” it said.

In May 2017, IA “convened a cadre of boosters to retroactively address the ‘copyright uncertainty’ around its practices,” said the brief. That resulted in a 2018 white paper and statement coining the phrase “controlled digital lending,” it said. The most critical component of CDL is maintaining the “owned to loaned ratio": that libraries in possession of physical books can make and distribute their own unauthorized ebook copies, but only after implementing measures to ensure that the physical book and the ebook never circulate at the same time, said the brief.

In 2018, IA implemented the Open Libraries Project “and essentially abandoned even the pretense of maintaining the owned to loaned ratio,” said the brief. Under the program, IA uses an “overlap analysis” to “radically expand the number of ebook copies simultaneously available to users” on the website, it said. Partner libraries submit data listing all the ISBN numbers for books in their catalogs, which IA “runs against its list of millions of physical books it keeps in shipping containers,” the brief said. “Every time there is a match, IA permits one more copy of that ebook to be borrowed,” it said.

The overlap analysis “turned IA’s Website into a centralized hub that lends more books than it or any contributing library actually owns,” the brief said. IA acts as “a national entity that lends other libraries’ books," with a goal of building a “robust system to circulate the resulting e-books to millions, and eventually billions of people,” the brief said.

In March 2020, IA lifted all ebook lending caps on its website, launching the National Emergency Library (NEL) in response to the COVID-19 pandemic, the brief said. While it acknowledged that NEL “deviated from [CDL]), it defended the action by claiming that “'the number of concurrent borrows would never exceed’ the total number of library books nationwide that were inaccessible due to lockdown restrictions,” it said. Those libraries weren’t consulted, “and the Copyright Office promptly issued a letter raising doubts over the NEL’s legality,” the brief said.

IA abandoned the NEL on June 10, 2020, and reimposed lending caps, said the brief, but it “has no ceiling on the total number of concurrent copies it can lend, and can ensure instant access to all of its ebooks if it signs up enough partner libraries or obtains enough used books.” If the 9,000 public libraries in the U.S. became partner libraries and contributed copies of a particular popular book in their physical collection, "more than 9,000 users of the Website could read IA’s unlicensed ebook at once." it said.

IA’s website “funnels users” to Better World Books, a for-profit corporation chaired by IA chairman Brewster Kahle, said the brief. Kahle told a potential investor he hoped to buy 7 million additional books through BWB to pay for the digitization of books and to “provide funding back to the Internet Archive,” said the brief. Donate and purchase buttons are available on the IA website, and IA receives a payment every time a user buys a book from BWB, the brief said.

IA refuses to pay the license fees libraries pay and markets its website to potential partner libraries as a free alternative to authorized ebook lending, said the brief. It also “encourages libraries to populate their websites with links directing patrons to ebooks” on the IA website, “which some libraries have begun to do,” it said.

In his March 24, 2023, decision, U.S. District Judge John Koeltl for Southern New York granted publishers’ motion for summary judgment, saying “fair use does not allow … the mass reproduction and distribution of complete copyrighted works in a way that does not transform those works and that creates directly competing substitutes for the original,” the brief noted.

In the decision, Koeltl “balanced the interests required by copyright law – including ‘the public benefits IA’s copying will likely produce,'” concluding that any such benefits “were outweighed by the harm” of depriving publishers of revenue they were entitled to, said the brief. The analysis applied “even more forcefully to the [NEL],” the brief said. The court’s Aug. 11 stipulation permanently enjoins IA from engaging in any form of CDL with respect to the publishers’ works and any other titles published by them in ebook form, it said. The 2nd Circuit should affirm, the brief said.