Low quality books that appear to be AI generated are making their way into public libraries via their digital catalogs, forcing librarians who are already understaffed to either sort through a functionally infinite number of books to determine what is written by humans and what is generated by AI, or to spend taxpayer dollars to provide patrons with information they don’t realize is AI-generated.
Public libraries primarily use two companies to manage and lend ebooks: Hoopla and OverDrive, the latter of which people may know from its borrowing app, Libby. Both companies have a variety of payment options for libraries, but generally libraries get access to the companies’ catalog of books and pay for customers to be able to borrow that book, with different books having different licenses and prices. A key difference is that with OverDrive, librarians can pick and choose which books in OverDrive’s catalog they want to give their customers the option of borrowing. With Hoopla, librarians have to opt into Hoopla’s entire catalog, then pay for whatever their customers choose to borrow from that catalog. The only way librarians can limit what Hoopla books their customers can borrow is by setting a limit on the price of books. For example, a library can use Hoopla but make it so their customers can only borrow books that cost the library $5 per use.
On one hand, Hoopla’s gigantic catalog, which includes ebooks, audio books, and movies, is a selling point because it gives librarians access to more for cheaper price. On the other hand, making librarians buy into the entire catalog means that a customer looking for a book about how to diet for a healthier liver might end up borrowing Fatty Liver Diet Cookbook: 2000 Days of Simple and Flavorful Recipes for a Revitalized Liver. The book was authored by Magda Tangy, who has no online footprint, and who has an AI-generated profile picture on Amazon, where her books are also for sale. Note the earring that is only on one ear and seems slightly deformed. A spokesperson for deepfake detection company Reality Defender said that according to their platform, the headshot is 85 percent likely to be AI-generated.
The book, which consists mostly of recipes and includes a disclaimer that the author and publisher “are not experts in the discussed topics,” didn’t include any dangerous information as far as I could tell, but appears to be AI-generated based on its stilted prose and formatting, consisting of summaries broken up with bolded subheadings common in outputs from LLMs like ChatGPT.
“Are these cookbooks written or reviewed by a dietitian or medical professional? Could a gastric bypass or cancer patient receive cooking instructions to make a meal contraindicated for their medical condition? If I were choosing for a library, I’d vet each one. With Hoopla, they are all there. Some might be excellent. Some might be dangerous,” Michael Blackwell, director of the St Mary’s County Library in Leonardtown, MD, told me in an email.
It is impossible to say exactly how many AI-generated books are included in Hoopla’s catalog, but books that appeared to be AI-generated were not hard to find for most of the search terms I tried on the platform. There’s a book about AI Monetization of Your Faceless YouTube Channel, or “AI Moniiziization,” as it says on its AI-generated cover. Searching for “Elon Musk” led me to this book for “inspiring quotes, fun facts, fascinating trivia, and surprising insights of the technoking.” The book’s cover is AI-generated, its content also appears to be AI-generated, and it was authored by Bill Tarino, another author with no real online footprint who has written around 40 books in the past year about a wide range of subjects including Taylor Swift, emotional intelligence, horror novels, and practical home security.
“I doubt that most library users of Hoopla are aware that some titles may be AI generated or unedited and of dubious quality,” Blackwell said. “They may assume that if the library is offering the information, it can be trusted. That is a problem.”
I also found at least one book on Hoopla that elsewhere admitted it was AI-generated: The Unknown Guest, a mystery novel by Rylie Dark, which on the United States Copyright Office is listed as “text generated by artificial intelligence.”
“Investigating these authors, their book covers, their social media, etc takes A LOT OF TIME, especially with the volume of questionable material increasing month to month (and that's not including the sheer amount of legitimate books published each month in adult fiction that I'm looking at),” one librarian who asked to remain anonymous so she could talk openly about her job, told me. “Is it the best use of my time doing this work on top of my other duties when customers may or may not care? And with the rising multitudes of AI generated content, will there come a point where it just ‘is what it is?’”
“My library, like most, does not have the resources to be checking Hoopla on a weekly basis to weed out what we wouldn’t want there,” Blackwell said. “Hoopla is marketed to libraries and should offer material of a quality that libraries want since we are not involved in the selection process, as we are when choosing from other ebook vendors or print.”
This type of low quality, AI generated content, is what we at 404 Media and others have come to call AI slop. Librarians, whose job it is in part to curate what books their community can access, have been dealing with similar problems in the publishing industry for years, and have a different name for it: vendor slurry. While the term now encompasses what seems like AI-generated content as well, it predates the rise of generative AI, and also refers to the glut of low quality, often self-published ebooks or book “summaries” that are common on Hoopla. As some librarians told me, the sheer quantity of books in Hoopla’s service makes it seem more valuable because it offers such a large number of books, but in reality that number is misleadingly inflated by this slurry.
“What hoopla is offering is this gargantuan amount of books. In actuality, it's padded with this slurry of poor quality materials that people likely don't want at all,” Luca Bartlomiejczyk, a librarian at Edith Wheeler Memorial Library in Monroe, CT, told me. “If you're going to say, ‘we have 15,000 ebooks on our platform,’ and 5,000 of those are low quality, AI generated or stuff that's just put on there without any kind of like oversight or selection criteria being followed, what are you actually offering to us?”
Several librarians pointed me to IRB Media, a publisher with hundreds of books on Hoopla, all of which are seemingly AI-generated summaries of other books. Theoretically there’s value in these summaries even if they are AI-generated—the popular CliffNotes brand which offers summaries of books did the same thing long before generative AI—but the sheer amount of summaries from publishers like IRB Media tend to show up in many search results, making it harder for people to find what they are looking for.
“If a patron, for example, wants to take out a copy of The Women by Kristin Hannah and what they find is a summary of The Women by Kristin Hannah with the word ‘summary’ written in a really tiny font, or they just think that it's the thing that they're looking for, and they don't look close enough then it's costing us the money for them to take it out,” Bartlomiejczyk told me. “It's costing them the time, and they're disappointed.”
In February 2022, two organizations of librarians, Library Futures and Library Freedom Project, sent Hoopla and OverDrive a letter demanding accountability over “fascist propaganda” materials being included in their services.
“These are books that are of such low accuracy and quality that not even Amazon will sell them. This includes materials from white nationalist publishers Arktos Media, Antelope Hill Publishing, and Castle Hill Services,” the organizations wrote. “In Hoopla, for example, the third search result for the word ‘Holocaust’ returned a Holocaust denial text not carried by most book distributors.”
In March, Hoopla Digital CEO Jeff Jankowski replied in a statement that said the books highlighted in the letter were removed and attempted to explain why they were included to begin with.
“The titles from these five independent publishers came to us from our network of more than 18,000 unique publishers,” Jankowski said. “They were added within the most recent twelve months and, unfortunately, they made it through our protocols that include both human and system-driven reviews and screening. As a result, we have taken immediate steps to improve our process.”
The librarian organizations said that they “appreciated” the response and that Hoopla removed the titles, but said the response was “insufficient.”
“Our questions remain about how Hoopla selects and approves materials for their collections. There is still a great deal of disinformation to be found on Hoopla,” they said. “For instance, when you search for ebooks about ‘homosexuality’ and ‘abortion,’ instead of factual informational content, the search results are largely self-published religious texts designed to misinform and scare library readers about sexual and reproductive topics.”
Several of the librarians I talked to said that they are worried about discussing this problem because of the growing hostility towards libraries and groups like Moms for Liberty demanding that books about LGBTQ rights, race, and ethnicity be removed from libraries. One the one hand, librarians want to curate their collections and make sure their patrons are getting access to quality information. On the other hand, they don’t want people to think that they are trying to censor what materials patrons can access in way that’s comparable to what organizations like Moms for Liberty want.
“It's this really fine line. This is not a request to censor materials. This is a request to ask companies that sell materials to the public with taxpayer money to be accountable for what is on their service, and to be transparent,” Jennie Halperin, executive director at Library Futures told me.
“We're always standing directly against censorship at every step of the way. That's what librarians do,” Bartlomiejczyk told me. “It's part of our code of ethics that we all follow, but professionals trained in best practices in selection of materials, i.e. librarians, to make these decisions would make a huge difference in the stuff that's being selected for these platforms.”
Hoopla and OverDrive did not respond to a request for comment.
None of the librarians I talked to suggested the AI-generated content needed to be banned from Hoopla and libraries only because it is AI-generated. It might have its place, but it needs to be clearly labeled, and more importantly, provide borrowers with quality information. Much like Wikipedia editors, librarians are curators of information for society. Wikipedia editors are unpaid volunteers. Librarians told me they were already understaffed, overworked, and under attack by conservative groups before AI-generated content became an issue. Both groups now face an entirely new threat to their core mission: generative AI’s ability to create an infinite amount of low quality information. All librarians are asking for at this point is that Hoopla explain exactly how its selection process work, and hopefully improve it.
“Platforms like Hoopla should offer libraries the option to select or omit materials, including AI materials, in their collections,” Sarah Lamdan, deputy director of the American Library Association, told me. “AI books should be well-identified in library catalogs, so it is clear to readers that the books were not written by human authors. If library visitors choose to read AI eBooks, they should do so with the knowledge that the books are AI-generated."
