Télison Aurélien denies using artificial intelligence, while AI detection tools “suggest” the opposite, causing a stir in the literary world
Until a few days ago, Télison Aurélien was one of the major revelations of this year’s French-language literary scene. His first novel, *C’était ça ou Mourir* (“It Was Either That or Die”), received rave reviews, was a major commercial success, and was in the running for major literary prizes, including the Goncourt. Now, however, the 38-year-old author is facing an accusation that is causing a stir in the publishing world: that he used artificial intelligence (AI) to write his book.
The controversy has taken on significant proportions precisely because of the novel’s impressive success. The book recounts the odyssey of a history and geography teacher who flees Haiti due to violence and attempts to reach Canada. The previously unknown poet has seen his first novel praised by the media—including *Le Monde*—and sell over 40,000 copies in France. He has already won the Fnac Novel Prize and is a contender for the Goncourt Prize.
ΣΧΕΤΙΚΑ ΑΡΘΡΑ
The clues that sparked the debate
As reported by the Wall Street Journal, the controversy began on the “Balance ton Claude” account on X, which claimed to have run excerpts from the book through Pangram, a tool designed to detect AI-generated text. After the results were made public, academics specializing in artificial intelligence conducted their own tests.
Kevin Bryan, a professor at the University of Toronto, stated that all the excerpts he examined were flagged by Pangram as 100% AI-generated. Tuhin Chakrabarti, a professor of computer science at Stony Brook University, checked the entire novel using the same tool and reported that 92% of the text was flagged as the product of artificial intelligence.
These results, however, do not constitute proof on their own. The reliability of AI detectors remains a subject of debate, as there is a risk that a human-written text could be incorrectly identified as machine-generated. Pangram claims that, based on internal testing, its latest model produces one false positive result per 24,000 cases.
The author’s response
Orélien categorically rejects the allegations and argues that the writing characteristics deemed suspicious stem from the literary and oral traditions of Haiti and the Caribbean, where repetition, rhythm, and vivid imagery play a central role.
“People can discuss my style, but turning its techniques into evidence of artificial intelligence has nothing to do with literary criticism,” he said.
The French publishing house Grasset also sided with him, rejecting claims that the work was written or created using AI. The president of the Académie Goncourt, Philippe Claudel, also expressed skepticism regarding the allegations, noting that the evidence presented so far is insufficient to condemn the author.
The case has also sparked a second controversy. Supporters of Aurélien argue that this is disproportionate scrutiny targeting a Black author who writes about immigration, while others see the allegations as an attempt to exclude him from the race for major literary prizes. Le Monde notes that the next critical juncture will be the second round of nominations for the Goncourt Prize on October 6.
From Haiti to the Top of the Bestseller Lists
“C’était ça ou Mourir” follows Jonas Dorléon, a professor who flees Haiti due to gang violence and travels through South America and the U.S. on his way to Canada. Aurélien, who settled in Montreal after the devastating 2010 earthquake in Haiti, has stated that he completed the first draft of the novel in 2019.
The book’s success has already spread beyond France’s borders. The rights have been sold in more than 20 countries, and the English translation is scheduled for release in the U.S. in 2028.
How reliable are AI detectors?
The controversy surrounding Aurélien highlights a broader issue: how safe is it to use AI detectors as evidence of a text’s authorship? Although tools such as Pangram boast high accuracy rates according to their creators, their results are not considered infallible.
To date, similar cases have mainly involved commercial fiction. The case of *C’était ça ou Mourir* is different, as it involves a book that has been critically acclaimed, is a bestseller, and is in the running for one of the most prestigious literary awards.
That is why the issue now goes beyond the question of whether or not this particular author used AI. It touches on a much broader discussion about the future of literature: what happens when we can no longer be certain that a acclaimed text was written exclusively by a human—and to what extent the literary world is ready to face this new reality.
Source: in.gr
