Amazon has removed another spurious book after American scientist Melanie Mitchell recently discovered an AI-generated imitation of her 2019 work, “Artificial Intelligence: A Guide for Thinking Humans” on the platform. This incident, part of a growing trend of low-quality AI-generated books on the e-commerce site, has sparked increasing concern and frustration among authors.

The rise of AI-generated imitations

Mitchell, who is a Davis Professor of Complexity at the Santa Fe Institute, found her work copied in a 45-page book, apparently filled with awkward phrases and listed under the author “Shumaila Majid.” This book, appearing alongside Mitchell’s original work on Amazon, raised suspicions due to Majid’s lack of an online presence and the proliferation of similar books under the same name. The profile has since been taken down.

Confirming Mitchell’s suspicions, WIRED consulted the deepfake-detection startup Reality Defender, whose analysis deemed the book 99% likely to be AI-generated. Speaking to the tech news organisation, she said, it “made me mad.”

“It’s just horrifying how people are getting suckered into buying these books,” she added.

Tech giants grapple with quality control

Following the inquiry, an Amazon spokesperson confirmed to How To Be Books: “While we allow AI-generated content, we don’t allow AI-generated content that violates our Kindle Direct Publishing content guidelines, including content that creates a disappointing customer experience. All books in the store must adhere to those content guidelines, including compliance with intellectual property rights and all other applicable laws.”

Read: Rory Cellan-Jones calls Amazon AI-generated bio ‘utter tosh’

Apparently, the issue extends beyond Mitchell. WIRED reported that Fei-Fei Li, a pioneering computer scientist, found over a dozen summaries of her memoir “The Worlds I See: Curiosity, Exploration, and Discovery in the Age of AI” on Amazon, many of which appear to be AI-generated. Reality Defender’s analysis of one such summary revealed a complete rewrite, likened to a “KidzBop version of the real thing.”

While the summary version has been taken down, an entry still appears on Goodreads, which is owned by the tech giant, and it has received a two-star rating. There appears to be no evidence that the publishing house “Kilback Publishing” actually exists. What’s more is there are 22 books of the same nature, including Omid Scobie’s book “Endgame,” a summary of Timothy Ferriss’ work “The 4 Hour Body,” and much more. The profile does not seem to appear on Amazon, but there are some dubious versions of the account such as “Miracle Kilback Publishing.”

Notable writer Jane Friedman and BBC journalist Rory Cellan-Jones have also fallen victim to unauthorised AI-generated publications. In response to these incidents, Amazon has taken down the books and stated its intention to implement measures to stem the flow of AI-generated content that fails to meet its guidelines.

Read: Amazon halts AI-generated books impersonating author Jane Friedman

Mitchell, reflecting on the legality of the situation, expressed doubts about the possibility of legal action, citing the inability to copyright book titles. Despite this, she remains concerned about the implications of AI-generated imitations on intellectual property and reader trust.

The ongoing struggle with AI-generated content highlights the challenges facing digital platforms in maintaining quality and authenticity nowadays. It will certainly be difficult to stem the flow of these books, especially if AI content is generally accepted on to the self-publishing site. There is undoubtedly a high level of irony in AI specialists being plagiarised by the very technology they’re writing about.