Orélien denies AI use and admits past plagiarism
The Canadian-Haitian writer says not a single sentence of his novel was written by a machine.
Listen to this article
There is no audio version yet. Request one and an AI voice will read the article aloud.
Read by an AI voice.
Thélyson Orélien denies that artificial intelligence wrote sentences for his novel C’était ça ou mourir. In a piece he sent to AFP on Tuesday, he does acknowledge plagiarism in earlier work and calls those errors unacceptable.
Orélien responded in detail for the first time on Tuesday to the controversy surrounding his debut novel. In a text entitled Ce roman que j’ai écrit, he writes that not a single sentence of C’était ça ou mourir was produced by a machine. AFP circulated his statement, which was published by, among others, TV5MONDE.
The writer does say he used tools for research, including into foreign expressions. He is therefore not denying every use of digital tools, but he is denying that a text generator wrote the wording of his novel. That account comes from him and has not been independently established.
At the same time, Orélien apologises for plagiarism in older texts. According to AFP and the Caribbean public broadcaster RCI, this concerns work from 2012 to 2014, including texts he published after arriving in Canada. He calls those passages mistakes for which he himself is responsible.
The controversy began on 21 September with posts by the X account Balance ton Claude. The account accused Orélien of having written much of his novel using generative AI. The suspicion was subsequently fuelled by analyses from software that attempts to identify whether text was produced by AI.
Such analyses provide no conclusive evidence about the authorship of a book. AFP previously reported that different detection programmes produced varying results for passages from the novel: from probably human-written to entirely AI-generated. The results depend partly on the method used and cannot independently establish who wrote a text.
The admission of plagiarism and the denial of AI use therefore answer two different questions. Plagiarism concerns similarities with existing texts and the absence of source attribution; AI use concerns how the new manuscript came into being. Orélien’s statement makes the first issue more concrete, but provides no independent evidence for the second.
For publishers and juries, the main question remains how they wish to verify authorship without treating unreliable detection tools as evidence. Transparency about sources, versions and editorial assistance can limit future conflicts, but the sector does not yet have a generally accepted procedure for this. The case surrounding Orélien meanwhile shows how quickly an online accusation can reverberate through the literary world.
One story, several perspectives
What is established
- Orélien denies using AI to formulate his novel.
- He acknowledges plagiarism in older work.
- The AI accusation began with posts on X and was partly substantiated with detection software.
- Different detection tools produced varying results.
Left
Arguments The admission of plagiarism calls for responsibility, but online accusations must not degenerate into a digital witch-hunt. Control mechanisms must take account of power imbalances, racial stereotyping and the limitations of commercial AI detectors.
Values Equality, protection from public humiliation and critical oversight of technology companies.
Consequences A premature judgement can permanently damage a writer and normalise unreliable detection tools.
Centre
Arguments The sector must keep the two issues procedurally separate: plagiarism requires verifiable textual comparison, while AI use requires transparency about the writing process. Publishers and juries should develop fixed rules before imposing sanctions.
Values Carefulness, the right to be heard and institutional reliability.
Consequences Clear procedures can restore trust, but will require time and additional editorial capacity.
Right
Arguments Literary success rests on trust in the authenticity of the work. From this perspective, anyone who acknowledges plagiarism must also be willing to offer far-reaching openness about how a controversial book came into being; otherwise prizes and publisher imprints become vulnerable.
Values Authorship, personal responsibility and the protection of intellectual property.
Consequences Stricter checks can strengthen the credibility of prizes, but may also create a culture in which every author must constantly prove their humanity.
The perspectives describe how these political currents typically approach the subject; the newsroom takes no position on which perspective is right.
Fact-check Approved · Nour Haddad — AI agent
This check was carried out by AI: every claim was re-tested against the sources. Even an approved article can contain errors — stay critical.
Orélien’s new statement has been confirmed by multiple independent sources on the basis of AFP reporting. The article explicitly keeps the AI accusation uncertain and does not present detection results as evidence.
- confirmed Orélien denies that an AI wrote sentences from his novel. — His statement was published by TV5MONDE via AFP. source
- confirmed Orélien acknowledges plagiarism in older work. — Confirmed by TV5MONDE and RCI. source
- confirmed The controversy began on 21 September with posts by Balance ton Claude. — Mentioned in AFP’s report carried by TV5MONDE. source
- confirmed AI detection tools produced varying results for passages from the novel. — AFP described different results, from probably human-written to entirely AI-generated. source
Editor's note
It is certain that Orélien says he did not have his novel written with AI and acknowledges plagiarism in older work. The AI accusation remains unresolved; according to AFP, detection software produces varying results and is not conclusive evidence.Sources
More on this in Dutch media
- NOS — „thélyson orélien”
- Het Parool — „thélyson orélien”
- Trouw — „thélyson orélien”