The Publishing Times

All the News That’s Fit to Publish

publishing-news · Guides

Thélyson Orélien Admits Past Plagiarism but Denies AI Wrote His Goncourt-Dropped Novel

In a statement to AFP, the Quebec author apologised for plagiarism in the 2010s but said no sentence of C'était ça ou mourir was created by a machine.

What did Orélien say? Thélyson Orélien, the Quebec-based author whose debut novel was dropped from the Prix Goncourt selection over suspected AI use, says no machine wrote his book. In a statement sent to the AFP news agency on Tuesday, October 13, 2026, and reported by franceinfo, he insisted that "no sentence" of C'était ça ou mourir "was created" by artificial intelligence.

In the same statement he also admitted that he "committed unacceptable errors" by plagiarizing the work of established writers in the 2010s.

It is the first time he has spoken at length since he cut short his French promotional tour on September 27 and went back to Quebec, where, franceinfo says, he had stayed silent since.

"A calumny and a heresy"

The statement is titled "Ce roman que j'ai écrit," or "This novel that I wrote." Orélien calls the claim that the book was written by AI "a calumny and a heresy." He says he used tools "here and there" for research, especially for foreign idioms. "But no sentence of this book was created by a machine," he wrote.

He says he began the novel "in autumn 2017, on the corner of a table" in Quebec, where he has lived since leaving Haiti in 2010, and finished a first draft in 2019. Franceinfo notes that he did not provide anything to prove that timeline.

"I wrote this novel myself," he said. "It comes from my wounds and those of other migrants. That wound of exile that never heals." C'était ça ou mourir is published in France by Grasset.

The plagiarism admission

The apology covers his early years in Canada. "I was a little over 20, I was probably in too much of a hurry, without really measuring what I still had to learn. I wanted to write, I read a lot," he wrote. "But admiring does not authorize taking, attributing to myself texts that are not mine." He added: "I regret it deeply and apologize without reservation."

That history is what first knocked the book off course. On September 25, his Canadian publisher, Les Éditions du Boréal, suspended the novel's promotional tour after the Montreal daily La Presse reported that a column Orélien published in 2013 was almost entirely a lightly reworked translation of an article that had appeared days earlier in The Philippine Star.

The AI accusation

The AI claims came from essayist Samuel Fitoussi, co-creator of the "Balance ton Claude" account. Fitoussi has said he ran excerpts of the novel through Pangram, the American AI-detection tool, and that he raised the alarm because he fears that "after several millennia of civilization founded on the written word, AI will cause the disappearance of reading and writing."

The Académie Goncourt then removed the novel from its selection, saying the decision was driven by a wish to protect the integrity of the prize and of the Académie, whose job is to honour literature written by women and men. The Prix Renaudot went the other way and kept the book among the nine novels on its second list.

Orélien said criticism is always legitimate, "but the threats, the media lynching and the racist harassment I have been subjected to for weeks are unbearable."

The book has kept selling through all of it. Franceinfo reports sales of more than 65,000 copies.

Part of a bigger fight

The case has set off a wider argument in French publishing about AI and authorship, and franceinfo links it to the human-written certification schemes announced at the Frankfurt Book Fair this month.

It also lands the same week that Zimbabwean writer Simbarashe Kundizeza denied using AI to write his Caine Prize-winning story, another case where Pangram results sit at the centre of the accusation. Both writers deny the claims.

Why it matters for authors

Your past work can resurface. The plagiarism that undid this book's launch happened more than a decade earlier, in journalism, not fiction. Prizes and publishers now look at a writer's whole record.

A denial needs evidence. Orélien gave a timeline but, as franceinfo pointed out, no proof of it. Dated drafts, notes and file histories are what turn a denial into a defence. PT has a guide to keeping records of human authorship.

Prizes are setting their own rules. The Goncourt and the Renaudot looked at the same book and reached opposite decisions. Until the industry agrees on standards, each prize will draw its own line.

Sources