Goncourt Removes Orélien Novel Over AI and Plagiarism Claims
The Académie Goncourt has removed Thélyson Orélien's C'était ça ou mourir from its longlist for the Goncourt prize. The academy said the withdrawal was driven by a desire to preserve the integrity of the prize and the academy itself, according to reporting published 25 September. The Guardian
The decision ends, at least for this season, the prize path of a debut that had moved quickly through the French literary system. C'était ça ou mourir is Orélien's first novel. It was published in France by Grasset in August. It has sold about 35,000 copies, won the Prix du roman Fnac and sold translation rights to more than 20 countries. The book was in contention for the Prix Femina as well as the Prix Goncourt.
Orélien is 38. He was born in Haiti and is described as a Canadian-Haitian writer. His denial and the institutional response unfolded over four days. Allegations that the novel was written with artificial intelligence emerged on X on 21 September, the same day the book won the Fnac prize.
The initial claim came from an anonymous X account, Balance ton Claude. The account said it had fed the novel into Pangram, an AI detection software that estimates the likelihood text was machine-written, and found it almost entirely written by artificial intelligence. Le Figaro later identified the author of the account as its columnist Samuel Fitoussi.
Orélien denied using AI tools to write the book on Tuesday, a day after the allegation surfaced. He said the allegation was racially motivated. His Canadian and French publishers stood by him after the accusation. Reuters
Académie Goncourt president Philippe Claudel had urged caution over the AI allegations on Wednesday. A separate allegation then widened the case. Libération alleged that Orélien plagiarised an award-winning short story from the preface of Louis Pauwels' The Morning of the Magicians, originally published in 1960.
The academy's statement cited both strands. It pointed to past plagiarisms by the author and to analyses by researchers and journalists indicating the book was very largely the product of artificial intelligence.
The broader context here is institutional rather than purely technological. Prizes like the Goncourt act as arbiters of literary standing and as market makers. A longlisting confers sales, translation contracts and a place in school curricula and backlists. Withdrawal is rare. It protects the jury from judging a text whose origin is in doubt. It also shifts judgment from literary merit to forensic questions about process that juries are not equipped to settle alone.
Looking at what this means for publishers and prize administrators, the sequence exposes a procedural gap. Detection software provided the spark, journalistic comparison supplied a second charge of plagiarism, and denial framed the dispute in terms of discrimination. None of those registers resolves the other. AI detectors give probabilities, plagiarism analysis turns on textual comparison, and questions of racial bias concern motive and reception. The academy has now treated the accumulation of doubt as sufficient to act, without publishing its own evidentiary standard.
Looking ahead to what may come next for writers working between languages, histories and markets, the stakes extend beyond one title. Orélien entered as a first novelist with rapid commercial and critical uptake. Grasset and his Canadian publisher defend an author whose book continues to sell, while the Goncourt and potentially the Femina must decide how far forensic scrutiny should reach into selection. The next test will be whether publishers, prize bodies or the author produce verifiable drafts, notes or editorial records, and whether French literary institutions codify rules on AI assistance before the shortlists narrow further.
