Even AI detection companies like Pangram got in on the action, claiming the book was primarily AI-generated.
Still, Hachette appears to be moving forward with plans for a U.S. release later this year.
Yesterday, The New York Times published an independent study in which it said it used “several AI detection tools to analyze passages of novels and found recurring patterns characteristic of AI-generated texts, including gaps in logic, overuse of melodramatic adjectives, and overreliance on the rule of three.”
That’s it. Hachette suspended publication in the UK and canceled its planned US debut. Late last night, the Times received a comment from author Ballard denying that he had used AI in writing his novel. Still, Ballard added that it’s possible that a friend who helped edit the book used AI.
Ballard added: “This controversy has changed my life in so many ways, my mental state is at an all-time low, and my reputation has been destroyed for something I did not personally do.” Ballard added that she is taking legal action.
But what if it’s “good enough”?
This was one of the first big AI controversies to hit the traditional trade publishing world, where old gatekeepers still largely ban AI, at least when it comes to drafts. (Synopsis, editorial feedback, plot suggestions…all of this is much more vague.)
Regardless of what actually happens in this situation, a similar disruptive pattern to the one currently roiling the music industry is likely to occur in publishing, where tools like Suno are increasingly being used to generate songwriting demos and (at least in places like Spotify) completely AI-produced music. Many artists and even distributors like LANDR resist the use of such AI, but many ordinary people don’t care. Their view seems to be something like this: If music sounds good, or even good enough, what difference does it make where it comes from? And frankly, how different is ultra-glossy pop from what Suno produced?
in the case of shy girlmany readers enjoyed this book and even promoted it online, despite many claims that AI’s writing is the worst and easily identifiable. It may scare and frighten actual writers, but it’s still a reality they have to face.
