Major Publisher Drops Horror Novel ‘Shy Girl’ Amid AI Allegations

Last updated: 2026/03/21 at 4:11 PM

News Room Published 21 March 2026

If you’ve been on social media sites like TikTok, Instagram, or YouTube, or even Reddit, you’re probably no stranger to unintentionally stumbling into AI-generated content. But up until now, at least, it’s unlikely you’d expect quite the same experience at a mainstream bookstore.

Now Hachette, one of the world’s biggest publishers, has canceled the US release of a new horror novel amid claims that it was AI-generated. According to The New York Times, which first reported the news, it is the “first commercial novel from a major publishing house to be pulled over evidence of AI use.”

Hachette pulled the book after The Times got in touch about rumors that it was at least partially AI-generated, and it’s now been removed from Amazon and the publisher’s website. Hachette is part of what is known as the Big Five, a collective of publishers that control most of the world’s book sales.

The claims had been spreading on YouTube and Reddit for months, as well as on book review sites like Goodreads. One Reddit post pointed to how “almost every noun” in the book is accompanied by an adjective, or the book’s overuse of weather similes for actions.

The book’s author, Mia Ballard, has denied the claims, alleging that the quirks in the book’s text may be due to her editor’s use of AI, without her consent. Ballard told The Times: “This controversy has changed my life in many ways and my mental health is at an all-time low, and my name is ruined for something I didn’t even personally do.”

Recommended by Our Editors

Even if there haven’t been many confirmed instances of major publishers letting AI-generated works reach the final stage until now, AI-generated work has been a growing problem for smaller outlets like science fiction magazines since tools like OpenAI’s ChatGPT first went mainstream. This has, in some cases, even forced some magazines to pause submissions.

Detecting AI-based writing has so far proved as much an art as a science, despite claims from tools that they can detect instances with widespread accuracy. Teachers and students across the US have had to contend with false positives from AI writing detectors.