Authors shocked at AI-generated book replicas on Amazon

Writers, including Rory Cellan-Jones, are dealing with extensive book spamming, often featuring numerous fake titles uploaded within a single day

For any writer, including Rory Cellan-Jones, publishing a book signifies a significant milestone. “As with any author, I regularly keep an eye on Amazon,” he stated. “And then, something unexpected occurred.”

The ex-BBC technology correspondent penned a memoir exploring his family’s history. On Amazon, what grabbed his notice was a biography of himself, sporting an amateurish cover, written by an unfamiliar author.

“I found it peculiar – who’s crafting a biography about me?” Cellan-Jones conveyed to the Observer. “I’m practical; it’s already tough for me to sell books about my own life, let alone for others to do so.”

However, upon scrutinizing a few excerpts, Cellan-Jones found that he had fallen prey to someone attempting to exploit his memoir by publishing a book seemingly generated by artificial intelligence. This was part of a surge in AI-generated titles that surfaced after the advent of ChatGPT, enabling individuals to produce pages of text without the effort of writing it.

Amazon is essentially enabling the spread of book spam and exposing it to the person most frustrated by it—Rory Cellan-Jones, the author

Cellan-Jones’s book, “Ruskin Park: Sylvia, Me and the BBC,” narrates his finding of a shoebox filled with letters from his mother, unveiling her romantic involvement with his father. Meeting his father at the age of 23, the book also details how his mother spent most of her life in a one-bedroom flat in south London. Describing it as a tale “about growing up with a single mother and a father who wasn’t present.

In contrast, the book contending with his family’s investigative narrative was, in Cellan-Jones’s view, “pure fiction.” He commented, “There are sections portraying the Cellan-Jones family, an academic household gathered around the dining table … His father depicted as a benevolent academic, his mother as a teacher. It’s simply untrue.

Later, I received an email from Amazon with the content: ‘You might be interested in this.’ Their algorithm had determined that I would prefer this unrelated book, instead of recommending my carefully crafted work into which I’ve invested my heart and soul… Essentially, they are allowing the proliferation of book spam and promoting it to the very person who finds it most infuriating.

Amazon did eventually remove the counterfeit biography and other works by the unidentified author, but numerous comparable titles continue to evade the safeguards intended to filter out inferior books.

For book spammers, it has been relatively easy to release multiple titles in a single day through Amazon’s Kindle Direct Publishing (KDP) system, which enables authors to independently publish both ebooks and printed books.

Under the pseudonym “Steven Walryn,” an individual published over 30 books, mostly comprising nonsensical and redundant guides on using various camera brands, along with a couple of fantasy romance novels. In May, they released 15 of these books on the same day. Amazon removed these books last week.

Amazon could not furnish details on the number of books it prevents from being published or those that have been removed. In August, Jane Friedman, a publishing writer, compelled Amazon to take down five counterfeit titles falsely attributed to her, which appeared to be generated by AI.

Amazon is evidently facing notable challenges with the influx of AI-generated content on its platform, and it seems to be working to catch up,” commented Nicola Solomon, the CEO of the Society of Authors (SoA).

Recently, Amazon announced that creators of new releases through Kindle Direct Publishing (KDP) must indicate whether their books contain AI-generated content and are limited to publishing three books per day—a development applauded by Solomon.

“Nevertheless, these modest changes appear to be more about streamlining Amazon’s procedures than enhancing the experience for readers and human authors,” she observed. The SoA is urging Amazon to distinctly label AI-generated products and allow readers to filter out titles generated by AI.

Rashik Parmar, the Group Chief Executive of BCS, the chartered institute for IT, stressed the necessity of government intervention to mandate the incorporation of a digital watermark on AI-generated content, facilitating seamless tracking.

Parmar remarked, “The UK has a distinctive chance to take the lead in responsible computing,” as long as legislation supports British IT professionals in upholding a higher ethical standard. He also suggested the potential for “coded in Britain” to symbolize something extraordinary.

In reply, an Amazon representative mentioned that all publishers must comply with the company’s content guidelines. They emphasized, “We invest substantial time and resources to ensure adherence to our guidelines and remove books that violate them… While we allow AI-generated content, we prohibit AI-generated content that contravenes our content guidelines, particularly content that results in a negative customer experience.