FTC Investigates Publishing.com: Are AI-Generated Books a Real Problem for Amazon and Other Retailers?

With Business Insider reporting on the latest scandal concerning the AI-generated books question, it seems that the contest between artificial intelligence backers and traditional authors (and perhaps book retailers) continues unabated.

As Business Insider detailed, it appears that Publishing.com — a company which charges $1,995 to ostensibly teach clients how to publish an assortment of their own AI-generated books, and launch them on Amazon and other retailer platforms — is under the microscope of the Federal Trade Commission.

Several of those customers allegedly have wrestled with Publishing.com over high-pressure sales tactics and misleading claims.

“‘They constantly say to use credit cards, borrow money, put yourself in debt in order to afford this program,’ one customer wrote in a complaint. Others said that refunds were difficult or impossible to obtain,” Business Insider reported.

The issue speaks to a larger concern, however — Amazon is being flooded by AI-generated books, according to a plethora of reports from NPR, Esquire, and Today, among others. What is Amazon, and its competitors, doing to ensure that avid readers (and customers) are protected from low-quality AI-generated copy?

Amazon Does Not Explicitly Ban AI-Generated Works, But Does Have Guidelines

Via its own Kindle Direct Publishing page, Amazon outlines that works which involve AI-generated assets must be disclosed before proceeding.

“We define AI-generated content as text, images, or translations created by an AI-based tool. If you used an AI-based tool to create the actual content (whether text, images, or translations), it is considered ‘AI-generated,’ even if you applied substantial edits afterwards,” Amazon wrote.

It should be noted that Amazon does not explicitly ban or restrict works involving — in whole or in part — AI-generated assets. This has led to a variety of issues spawning from the proliferation of generative AI technology, as well as the profiteering involved in its deployment into the creative space.

As The Globe and Mail detailed, AI-generated authorship was the basis of a false biography profiling a victim of the Oct. 7 Hamas attack upon Kibbutz Nir Oz — this volume had to be manually reported to Amazon before being made unavailable. A book on the Canadian province of Manitoba’s mushrooms leaves out several entries on deadly fungi, and is believed to have been tossed together hastily by an AI model, as CTV News reported.

Today underscored the experiences of several authors who had seen their own works immediately accompanied by a bevy of scam workbooks and impersonated materials.

In response to the great volume of scam books, low-quality works, and other AI-generated offenders in the bookselling space, Amazon released the following statement.

“We both proactively prevent books from being listed as well as remove books that do not adhere to those guidelines, including content that creates a poor customer experience,” Amazon’s statement read, as cited by NPR.

“We have more recently begun limiting the publication of summaries and workbooks based on existing titles in our store.” It adds: “When patterns of abuse warrant it, we also suspend publisher accounts to prevent repeated abuse,” it continued.

Amazon was forced to implement a change limited the number of self-published books from an individual publisher utilizing its KDP platform to three per day. Of course, it is unlikely that many publishers utilizing that platform would organically achieve such a result, even if they were part of a larger collective or micropress — a policy which gestures solidly toward the issue of AI-generated works being pushed through the platform at a breakneck pace.

While it’s difficult to pin down exactly how much money Amazon — and other booksellers — are making from the sale of low-quality or otherwise less-than-authentic AI-generated books is hard to pin down. YouTube search results show several influencers making claims of earning thousands of dollars per month in residual income from their respective AI-authored portfolios, but without receipts, these claims may be self-promotional to some degree.

Generative AI Models Rely on Copywrite-Protected Content to Feed Upon: Legal Limbo

There’s also the issue of how generative AI often compiles its own foundational database from which all subsequent works are issued forth from.

“The scale of AI-generated content is so overwhelming in part because it’s so cheap to create. That’s because no one is paying for one of the highest-quality sources of training data: books. Investors and leaders in AI know that shelling out for this content would break their business; some have acknowledged as much,” Esquire noted.

“In a response to the U.S. Copyright Office, the venture capital firm Andreessen Horowitz wrote, ‘Imposing the cost of actual or potential copyright liability on the creators of AI models will either kill or significantly hamper their development,’” the outlet added.

As the industry matures, Amazon and other retailers in the bookselling space may find themselves under increased pressure to label, separate, or entirely divest themselves from AI-generated work. Or, on the other hand, new models free of these legal constraints — and operational constraints — may arise to satisfy the needs of a new generation of readers. The die, as yet, remains uncast.