In mid-June, the Internet Archive terminated their National Emergency Library (NEL)two weeks prematurely in response to a lawsuit brought by publishers Hachette, Penguin Random House, Wiley, and HarperCollins. Publishers were not the only parties concerned with the copyright issues raised by NEL. When the NEL launched in March, the Authors Guild issued a statement saying that they were “shocked that the Internet Archive would use the Covid-19 epidemic as an excuse to push copyright law further out to the edges, and in doing so, harm authors, many of whom are already struggling.” However, a more significant issue for copyright holders remains unresolved: the Open Library, and it’s “controlled digital lending” model. For visual artists, Open Library poses a particular threat.

The NEL was an extension of the Open Library, and permitted lenders immediate access to digital copies of millions of books in Open Library’s collection. NEL was intended to give students, academics, and basically anyone able to set up an Open Library account unfettered access to books during a period when public health measures had shuttered brick-and-mortar libraries. Open Library permits registered users to “borrow” up to five digitized or ebooks books for a two week period. In pivoting to the NEL , they did away with the waiting period, meaning that users wanting to access books which were already “borrowed” no longer needed to wait until the ebook had been “returned.”

Open Library is an initiative of the Internet Archive, and seeks to “make all the published works of humankind available to everyone in the world.” Their efforts towards that end have been prodigious: by their own account, they have scanned over 1.4 million books. As an example of their reach, their website was visited by between 11 and 12 million visitors in a recent two week period, during which over 400,000 books were “borrowed” from their collection.

The Internet Archive has made entire works of visual artists – complete photographs and illustrations published within a scanned Open Library offerings – available to bad actors in a usable JPEG format.

Controlled Digital Lending: Cutting Out the Copyright Holder

The process by which the library lends books is “controlled digital lending” (CDL), which resembles the licensed ebook lending model libraries currently employ. Under CDL, the Open Library scans donated, bought, or lent books and uploads them to their server. That ebook is copied for each online borrower requesting the book (up to five, according the the Open Library). The difference between CDL and how traditional libraries operate is that traditional libraries purchase licenses for ebook lending. Those licenses emulate the experience of borrowing physical books: only one digital copy at a time can be borrowed, and that the borrower can’t access that digital copy after their borrowing time has ended (unless they borrow the book again). More importantly, the licenses negotiated by traditional libraries ensure that publishers and copyright holders control where and how their works are distributed, and are renumerated.

CDL, however, has written the copyright holders out of the equation. The books which Open Library scans have been donated or lent by libraries and users, acquired from closed libraries or library discards, or purchased. No effort appears to have been made to limit the contributions to books which are out of copyright. These books are scanned and uploaded to the Library’s servers, and converted to audiobooks, without obtaining permission from the copyright holders. In contrast, in ebook lending as exercised by traditional libraries, publishers and authors retain control of their works. They decide whether their books will be digitized and when, and whether they will be converted to audio books. Copyright holders derive an income from those additional licenses for ebook and audio book publications. Publishers and copyright holders also negotiate with libraries the licenses for digital lending of those ebooks and audio books.

Of Particular Concern to Visual Artists

The API created by Internet Archive has been already utilized to selectively search out and download all book page images which contain illustration, graphics, or photographs. In 2014, Python programmer Stephen Krewson published an article on how to extract only book page images containing pictures from the digitized books on Internet Archive. (Krewson falsely identifies the books as “public domain – most of the volumes on Open Library are still under copyright.) Essentially, Internet Archive’s API provides determined programmers a way to scan through the millions of book pages and snarf up all pages containing pictures and graphics. This would enable bad actors to create and monetize libraries of digital images. The copyright holders and creators of these works for the most part wouldn’t even be aware that their works have been made vulnerable by the Internet Archive’s practices.

Even without the API, visual artists are compromised by Open Library’s lending model. When accessing a book, the entire book – the JPEGs of each page – are downloaded to the user’s hard drive. “Returning” the book doesn’t delete that collection of JPEGs, which remains on the borrower’s hard drive, and can be accessed by anyone capable of accessing recent files. In fact, from NWU’s tests, Open Library deems a book “returned” at the end of the two-week borrow period, whether or not the borrower clicks onto “return” button.

This means that visual artists – illustrators, photographers, graphic designers, etc. – are particularly vulnerable. The Internet Archive has made their entire work, published within a scanned Open Library offering, available to bad actors in a usable JPEG format. There is little to stop someone from utilizing a script as described above, or scan through the JPEGs stashed on their computer, to create entire libraries of photographs and illustrations.