French novelist accused of using AI — MyJoyOnline
In France, Canadian-Haitian writer Telison Aurélien has been accused of using artificial intelligence to write his debut novel, C’était ça ou mourir. The Balance ton Claude group claimed that the book, for which the author has already received the Prix du Roman Fnac, was allegedly created almost entirely by AI.
Group’s statement and author’s response
As MyJoyOnline reports, members of the little-known Balance ton Claude group wrote on X that they checked several excerpts of the manuscript using the Pangram program. According to them, every check returned a result of “100 percent AI.” The group’s membership is unknown, while its representatives claim that the tool has an almost minimal probability of error.
The 38-year-old Aurélien rejected the accusations. In a comment to the newspaper Libération, he said that he had always loved creativity and literature and saw no reason to use a machine. The writer also said that he and his publisher were preparing materials to confirm that he wrote the novel independently.
Publisher Grasset defended the author, calling the accusations deeply offensive and stressing the originality and strength of his work. Aurélien explained that the imagery, repetitions and rhythms in the work are connected to his Haitian and Caribbean traditions.
More current news is available on the UA.News Telegram channel Telegram.
Awards and the reliability of detectors
The novel tells the story of a teacher who flees gang violence in Haiti and eventually reaches Canada, where he regains the opportunity to live without fear. Aurélien is also a finalist for the Goncourt and Renaudot prizes, the results of which are due to be announced soon. The Canadian publisher called the book its best-selling title in 60 years, and the novel is planned to be translated into about twenty languages.
Aurélien was born in 1988 in the Haitian city of Gonaïves and moved to Montreal in 2010. He claims that he wrote the first version of the novel in 2019, before text-generation programs became widely available.
Critics of Pangram and other AI detectors note that such tools cannot establish who exactly created a text. They analyze statistical and linguistic patterns that may be characteristic of both AI texts and human works. According to research, detectors may mistakenly label human texts as AI-generated or fail to recognize materials created by artificial intelligence and subsequently edited or paraphrased.
