Authors seem to be facing a new threat from artificial intelligence, with one finding books she didn’t write being sold by Amazon under her name. There are fears that ready access to generative AI tools could make it easy for people to impersonate writers without their permission.
The issue was . She found that five books purporting to be written by her were listed for sale on Amazon, with links to her Amazon author page – a connection she says has hoodwinked her fans into thinking they are buying legitimate books she wrote.
A reader first made Friedman aware of the fraudulent books after buying two and realising they were not up to the same standard as her previous work. “It’s quite a violation to see garbage published under your name,” says Friedman.
Advertisement
Although Friedman hasn’t been able to confirm that the books were created by an AI, she says she has “pretty good reason to believe” they are. She says their opening pages are similar in style to the content produced by generative AIs such as ChatGPT when she asked it to recreate her style.
Friedman reported the books to Amazon. In a , she says that Amazon initially asked her to provide the retailer with “any trademark registration numbers that relate to your claim”. She was unable to do this because she hasn’t trademarked her name.
“The systems are completely inadequate to deal with this,” she says. After żěè¶ĚĘÓƵ approached Amazon to comment on Friedman’s blog post, the titles claiming to be written by her were removed by Amazon.
“We have clear content guidelines governing which books can be listed for sale and promptly investigate any book when a concern is raised. We welcome author feedback and work directly with authors to address any issues. We invest heavily to provide a trustworthy shopping experience and to protect customers and authors from misuse of our services,” says an Amazon spokesperson.
Author impersonation pre-dates the rise of widely available generative AI, as author and journalist found out after publishing a book in 2019 and discovering fraudulent books were being sold on Amazon under her name. She thinks the scammer was attempting to game the promotion algorithms used by both Amazon and Goodreads, a book cataloguing website owned by Amazon.
“For new books like mine, your publisher puts a lot of juice in the algorithm,” she says. “They put you on Goodreads with a bunch of giveaways. And a scammer can attach an identity to yours and use some of that for themselves.” The problem will only get worse now that fake books can be generated by AI with little effort, she says.
, a partner at New York law firm Leason Ellis, says that many of the authors she represents register their names and pen names as trademarks so they can compel companies like Amazon to remove fraudulent entries in precisely these circumstances.
“Not all sites have takedown procedures for trademark complaints,” says Emerson. “Of those that do, many, including Amazon, will only process notices pertaining to registered – as opposed to common law – trademarks.” A registered trademark requires registering your brand or name with a formal body, whereas common law trademarks are established simply by using that brand or name first.
For those who don’t register their names as trademarks, such as Rose and Friedman, there are still routes to removing content falsely claiming to represent an author’s work– but it is tricker. “Often, it’s a question of getting in touch with the right human being at the retailer,” says Emerson.
Emerson says that authors could pursue the individual behind the book – or even the platform that is hosting the fake listing – for common law trademark infringement and unfair competition, or potentially right of publicity violations.
However, “pursuing such claims can be difficult as a practical matter”, she says, simply because of the difficulty in identifying the perpetrator. Emerson also acknowledges that authors who depend on platforms like Amazon for most of their book sales may be wary of taking legal action against the retailer, in case it causes ramifications for future sales.
For Friedman, the issue only looks set to grow. “These companies have put no guard rails in place,” she says. “It’s going to be whack-a-mole for days, weeks and months or years until they actually have real policies that prevent misuse of AI.”
