Phoenix 24
A literary controversy exposes unresolved questions about authorship and transparency.
MONTREAL
Canadian publisher Les Éditions du Boréal has ended its relationship with Haitian-Canadian author Thélyson Orélien after an independent examination found evidence that artificial intelligence was used during the writing of his acclaimed novel C’était ça ou mourir. The decision marks a significant development in a controversy that has challenged traditional expectations of literary authorship and editorial trust.
In a statement issued on October 21, the Montreal-based publisher said newly confirmed evidence contradicted Orélien’s repeated assurances that he had not used artificial intelligence to create the book. Boréal announced that it would not reprint the novel or publish additional works by the author.
The publisher acknowledged an important limitation: although AI involvement had been established to its satisfaction, the extent of its contribution remained unknown. The findings did not determine whether artificial intelligence generated substantial passages or supported more limited stages of the writing process.
That distinction matters because AI-assisted editing, stylistic revision and the generation of original prose raise different questions about creative responsibility.
The controversy began after an account on X called Balance ton Claude accused Orélien of extensive AI use. The allegations attracted attention because the novel had achieved commercial success and recognition within the French-language literary world.
Published in Canada in March 2026, C’était ça ou mourir explores migration, exile and survival through the experience of a Haitian character seeking a new life in Canada. The book received literary recognition, including the Fnac Novel Prize, and had been considered for major French awards.
As scrutiny intensified, Orélien denied using AI to write the novel. He also acknowledged unrelated instances of plagiarism in earlier work and reported receiving racist harassment during the controversy.
Boréal initially defended the author, explaining that its own editorial process had not involved artificial intelligence and that it lacked conclusive evidence about how the original manuscript had been produced.
The publisher subsequently sought documentation of the novel’s development, including earlier drafts and records that might clarify its creation. When those efforts failed to resolve the allegations, Boréal commissioned an external examination of the available working versions.
The resulting assessment prompted the publisher to reverse its position.
The case illustrates a broader difficulty facing contemporary publishing. Editorial revisions have always involved collaboration, but generative AI introduces new uncertainty about the relationship between an author’s declared creative process and the text ultimately presented to readers.
Detecting machine-generated writing through stylistic characteristics alone is not sufficiently reliable to establish authorship. Manuscript histories, document records and corroborating evidence may provide stronger grounds for investigation, although they also require careful interpretation.
The controversy does not establish that every use of AI constitutes literary misconduct. The central issue in this case concerns the publisher’s finding that the technology was used despite repeated assurances to the contrary.
For publishers, the implications extend to contracts, disclosure expectations and procedures for verifying disputed manuscripts. For writers, they concern the boundaries between technological assistance and the intellectual responsibility attached to a published work.
Boréal’s decision resolves its commercial relationship with Orélien but leaves unanswered how extensively AI contributed to the novel.
The unresolved question is becoming increasingly relevant across creative industries: when technology participates in producing a work, what must audiences and publishers know about the process?
When Evidence Speaks, Narratives Change.