France’s most prestigious literary award found itself at the center of a sharp contradiction this week. The Prix Goncourt jury had removed a debut novel from its longlist over suspected AI use, citing the need to protect literature written by “women and men.” Then the prize winner stepped forward and said he had used AI too.
The two cases are not identical. But the contrast between them has opened a wider conversation about what literary institutions actually mean when they draw lines around artificial intelligence, and whether those lines are drawn consistently.
Three Uses, One Admission
Philippe Jaenada, who won the Prix Goncourt for L’Inconnue du quai de Javel, disclosed on French radio the day after receiving the award that he had turned to ChatGPT three times during the writing process. The uses were distinct in nature and intent.
The first was deliberate mischief: a single sentence generated by ChatGPT, inserted into the novel to “play with the reader.” Jaenada revealed which sentence it was only to the Goncourt jurors on the day of the award, adding that if the information leaked publicly, he would know its source. The second use was satirical: a passage of AI-generated text included specifically to mock ChatGPT, which had produced what Jaenada described as “completely absurd answers” about a technical detail. The third was practical: research assistance.
Jaenada, 62, framed his position clearly. He argued that AI cannot be stopped or ignored, and that the only responsible approach is to acknowledge it, learn to control it, and make it useful. He had made a similar argument before the award, comparing AI tools to special effects in filmmaking and suggesting that declaring a text “100% human-written” would soon become impossible. At the same time, he drew a firm line: AI, in his view, will never be capable of writing with “heart, soul, or memories.”
This is a nuanced position. It treats AI as a tool that a human author directs, rather than a system that generates work autonomously. Whether that distinction holds up under scrutiny depends on how much of the creative judgment remains with the human, and how much is delegated.
The Orélien Case: A Different Outcome
The contrast that has drawn the most attention involves Thélyson Orélien, a Canadian-Haitian writer whose debut novel C’était ça ou Mourir had been a standout of France’s rentrée littéraire, the concentrated publishing season between late August and early November. The book sold more than 35,000 copies and won the Prix du roman Fnac before journalist Samuel Fitoussi ran it through an AI-detection tool called Pangram and concluded that much of it had been machine-generated.
The Académie Goncourt removed the book from its longlist in September, stating that analyses by researchers and journalists showed it was “in all likelihood very largely the product of an artificial intelligence.” The academy said it wanted to send “a clear message” that it did not endorse AI-generated texts.
Orélien denied the allegations, calling them “slander and heresy” and insisting that “not a single sentence in this book was created by a machine.” In the same statement, he acknowledged plagiarism in earlier work. His Canadian publisher, Boréal, subsequently said new evidence showed AI had been used in the writing process, contrary to the author’s repeated assurances, and ended its relationship with him.
The two situations differ in important ways. Orélien denied AI use and was contradicted by his own publisher. Jaenada disclosed his use voluntarily and framed it as intentional and limited. The scale of AI involvement, if the allegations against Orélien were accurate, appears to have been far greater. These are real distinctions.
Still, the timing of Jaenada’s admission, one day after an award whose jury had expelled another writer for AI use, has prompted pointed commentary online. Some observers noted that Jaenada faced no public backlash or institutional consequences, while Orélien’s career was effectively ended by the same institution. The question of whether race and status shaped the different responses has been raised directly on social media, and it is not a question that dissolves simply because the two cases have different facts.
What Literary Institutions Are Actually Deciding
Here is what most coverage of this story misses: the Goncourt affair is not primarily about AI. It is about how institutions define authenticity, and who gets to set those definitions.
The Académie Goncourt’s stated position is that it exists to honor literature written by human beings. That is a coherent value. The difficulty is that “written by a human” is no longer a binary condition. Authors have always used tools, editors, research assistants, and reference materials. AI sits somewhere on that spectrum, and exactly where depends on how it is used.
Jaenada’s three uses span a wide range: one is a creative experiment, one is satire, one is research. None of them, by his account, involved delegating the authorial voice to a machine. If that account is accurate, his use of AI is closer to a novelist using a search engine than to one outsourcing prose generation. The Orélien case, if the publisher’s final statement is taken at face value, appears to have involved something more substantial.
Literary prizes will need to develop clearer frameworks for these distinctions, not because AI use is inherently disqualifying, but because “AI was involved” covers an enormous range of practices. Without that clarity, institutions risk applying rules inconsistently, and inconsistency in high-stakes decisions tends to reveal, and sometimes amplify, existing inequalities.
In Short
The Prix Goncourt winner admitted using ChatGPT three times while writing his novel, one day after the same prize had removed another book from its longlist over suspected AI use. The two cases differ in scale and in whether the author was honest about it. But together they expose a gap that literary institutions, and many others, have not yet closed: the difference between having a policy on AI and having a coherent, consistently applied one.
Based on reporting from The Guardian.