If you have noticed that the quality of novels has declined in recent years, it might be because the books you’re reading weren’t written by a human.
The publishing industry has found itself in a tricky position with people arguing about the merits and drawbacks of artificial intelligence in world of literature. One major publisher was forced to cancel a famous novel after finding out the author was likely a machine. A prestigious literary prize is under scrutiny after one winner’s story was flagged as AI. Further complicating the matter is the flood of self-published authors creating whole novels using artificial intelligence.
These developments have prompted many to speculate about the future of the publishing industry and how it might impact those who enjoy curling up with a nice thriller to unwind after a hard day. But the issue might also affect human writers seeking to make some extra cash by telling stories.
The industry was awash with debate after Hachette pulled the book “Shy Girl” by horror novelist Mia Ballard. The book was published in February 2025 and garnered massive popularity on social media before Hachette picked it up for the United Kingdom and planned a U.S. release.
However, someone using an AI detection tool analyzed the text and determined the book was 78 percent AI-generated, according to The New York Times. The article noted typical patterns found in AI writing, including gaps in logic and repetitive structures. The publisher conducted an investigation before canceling the U.S. release and discontinuing the U.K. edition.
Ballard denied using AI herself to write the novel. She said an editor who worked on her novel was responsible for the use of AI. “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,” she said.
‘Shy Girl’ author releases statement after her novel was pulled amid allegations that she used AI to write it:
— Pop Crave (@PopCrave) March 20, 2026
“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.” pic.twitter.com/hNXpDdkThK
Romance author Lena McDonald recently faced a similar backlash when those who read her book “Darkhollow Academy: Year 2” discovered an AI prompt that was accidentally left in the published text. The prompt showed instructions to rewrite a passage to match the style of another author.
McDonald faced a barrage of accusations of using AI to generate or heavily edit the novel. She apologized to her audience and explained that she had used AI as an editing tool because she could not afford a professional editor as a full-time teacher and mother. If her story is true, it shows how the advent of AI can affect even authors who don’t actually use it to generate the text of the book.
Lastly, we have the Commonwealth Short Story Prize, which became the latest flashpoint in the industry in May. A regional winner from Trinidad named Jamir Nazi published a short story titled “The Serpent in the Grove” in Granta magazine after winning the award in the Caribbean category.
However, after winning the award, AI detection tools flagged the story as 100 percent AI-generated. Other winners also fell under scrutiny after this was revealed.
The author defends his writing, saying he relies on speech-to-text due to health issues. Granta’s publisher also ran the story through an AI detection tool and reported that it said the piece was “almost certainly not produced unaided by a human.”
There will likely be plenty more scandals like this in the near future as the publishing industry adjusts to the prevalence of artificial intelligence. But further complicating the matter is the reality that AI detection tools have not proven effective at determining whether a piece is written by man or machine.
Multiple studies showed that these tools often generate false positives by incorrectly labeling human-written text as AI-generated. Error rates have reached double digits at times, depending on the tools and the author’s writing style. In fact, one detector developed by OpenAI only correctly identified AI usage in only 26 percent of samples while marking nine percent of human writing as machine generated.
This means that many authors could be wrongly accused of using AI because of faulty tools. To those experienced with reading both human and AI-generated texts, some of these examples are easy to suss out.
The debate also centers on how authors should use AI. Purists contend that no AI should be used at any part of the process. Others suggest that AI use is acceptable if the author discloses that they used it to write the book. Indeed, Amazon has begun requiring authors to note whether they used AI to generate the text of a book or even the images on the book cover.
Most seem to believe AI is fine for brainstorming, research, feedback, identifying inconsistencies, etc. But they also agree that authors should not be using the technology to write the book, or even parts of the book, for them.
As an avid reader of mysteries and thrillers I’ve noticed many books that are clearly written by artificial intelligence. The patterns can be subtle, but they are there. While it can write an entertaining enough story, it can’t mimic the human element, the parts where a good author can use prose to make us feel something.
Yet, some are still making money by using AI to write books. The advantage they have is that AI can churn out novels that could take months or even years for a human author to write. This allows people to publish books that might not sell as much (because they are terrible) but will sell enough to make a decent profit because of the volume. At the end of the day, it’s a numbers game.
The issue is that these books are starting to flood the market, making it harder for authors to connect with readers deluged by AI-written pieces they believe to be written by humans.
As someone who just published his first novel, it’s been a concern for me. But in the end, I think authenticity will win out. AI will certainly get better at sounding closer to human, but I doubt it’ll ever be able to mimic humanity well enough to compete with skilled human authors who know how to connect with other humans.
Still, it will likely be some time before the industry and companies like Amazon navigate these issues in a way that ensures the market is fair.
