To be artificial or not to be artificial? It now seems to be one of the central themes (and doubts) in the contemporary literary (and editorial) debate. Especially in France, where the Canadian-Haitian writer Thélyson Orélien has been excluded from the Goncourt prize (the most prestigious transalpine award, which will be awarded on November 3) because he would have made wide use of artificial intelligence in drafting his C’était ça ou mourir (Grasset). The conditional, as we will see, is mandatory.
The Accusation Against Thélyson Orélien and the Revelations
But let us proceed in order. Published in March in Canada and in August in France, the novel by the 38-year-old writer was accused of being written “almost entirely” with the support of AI by an anonymous account on X, which had subjected the text to the analysis of the software Pangram (a tool created precisely to detect texts generated artificially). If at first the Académie Goncourt suggested caution, in Canada the press revived some of Orélien’s stories accused of plagiarism.
And so, after research that offered no certain data (not a few maintain that the same analysis tools are ineffective, and others wonder if it makes sense for AI tools to render verdicts on texts written with or without AI support…), the Academy decided to exclude the novel with the aim of “preserving the integrity of the Prize”: “We neither approve nor tolerate in any way the creation of texts generated with the aid of artificial intelligence.”
Meanwhile, Samuel Fitoussi, a journalist from Le Figaro, said he was one of the people who manage the anonymous account from which the complaint originated. And here another branch of the controversy began.
The Many Reactions
Although the tour of presentations has been suspended, Orélien continues to defend his actions: “I have always loved creativity and writing, I don’t see why I should rely on a machine,” he said in an interview with Libération, and his publisher spoke of “malicious statements” against him. The comments are abundant, given that the news, which is dividing industry insiders, has been picked up by media around the world.
As for Thélyson Orélien’s book, which has already sold tens of thousands of copies abroad, it has been acquired in Italy by Rizzoli: “We appreciated it for its voice,” Massimo Turchetta, general director and publisher of the Mondadori Group, explained to Repubblica, adding: “The major publishing houses around the world participated in the international auction to acquire the rights. Boréal is delving into a textual analysis to clarify beyond any doubt what happened, and Rizzoli is waiting to have a fuller picture.”
In France, among others, the author who won the prestigious prize in 2021 Mohamed Mbougar Sarr spoke about the affair. The author of La plus secrète mémoire des hommes (e/o, Alberto Bracci Testasecca’s translation) spoke of the need to “not underestimate the fact that these attacks, doubts and questions are interwoven with a racism that comes from a political sphere that is easy to identify” and he continued to emphasize that “conservative accounts amplified the post and gave it a huge audience.”
Expanding the reflection, Sarr argued that “in the coming years the lawsuits of this kind will multiply against all authors, perhaps with growing distrust toward those who had not published before the explosion of LLMs” (large language models, ndr).
If just a few months ago Il Post ran an article with “tricks and sites that can give a hint” about a text’s humanity, the same Sarr reflects on the “dead end” we find ourselves in today: “Subjecting texts to an AI detector is another way to feed AI, to gorged capitalism’s cruelty in its techno-assassin version, and to make the machine not only an arbiter, but a God who will decide what is human and what is not human.”
“No book is written ‘by an artificial intelligence’”
As the debate inevitably broadened, it became clear how little is known about these language models and the use—creative?—of AI systems. Luca Sofri, founder of Il Post, on his blog Wittgenstein, reflected on the process by which a book comes to be written, with or without machine assistance: “[Orélien] didn’t wake up one morning, go to his desk, open the computer, and find a novel that had been written overnight by software without his knowledge.” At most, Sofri continued, one could say there was a use “consistent and instrumental to an idea and to a structured indication from the author of the book.” Of course, this does not deprive that an undisclosed use of such a tool could violate the rules of prizes and contracts with publishers.
All this once again raises questions about the creative capacities of AI, even in comparison with human abilities; “Do we decide that [the novel’s quality] is due to the wisdom and talent of the author,” or “do we conclude that a great book was produced by software following insignificant human instructions of little value”: and then perhaps we must overturn all that we have said to date.
“Writing with AI is perfectly fine”
And while in the United Kingdom there was talk of a label to mark works written by humans, it is worth reflecting on what it truly means to write using these tools. On Facebook, the editorial director of the magazine L’indiscreto, Francesco D’Isa, who often weighs in on these topics, reflected on the information overload generated by the use of artificial intelligence, since “it so greatly expands one’s cognitive reach that it sometimes overwhelms the brain.” And precisely commenting on the news of Thélyson Orélien’s exclusion from the Goncourt, he added: “Writing with AI is perfectly fine, indeed, if you work better using it I find it stupid not to use it…”
The same D’Isa later signed a reflection titled La caccia alle streghe contro chi usa l’IA per scrivere è stupida for the-bunker.it: “(…) I can use grandpa’s fountain pen, a typewriter, Word with autocorrect, an editor paid per page who rewrites half a chapter for me, or a language model that offers me twenty variants of a paragraph from which I save three lines. I may have read thousands of books and derived style and ideas from them, or I may not have read a single line. I may take a plot from a story heard at a bar or have it suggested by a dream; I may talk about real autobiographical facts or make them up entirely and call them true. The reader has every right to judge what they find on the page and to stop if they dislike it. The rest, however, is my business, and if you’re eager to uncover the truth, know that it will forever lie beyond your sight: you will never have any certainty. Writing hides the writer in a thousand ways and it is never, never—and I repeat never—a child of a single pen, even if you confine the writer in an isolation cell.”
Not the first case: from Shy Girl to Call Me, I’ll Hide the Body
If Orélien continues to make headlines also due to the book’s success, his case is not the first instance of a possible use of generative AI that has stirred debate in the publishing world.
Between late July and early August, for example, there was talk of Jerry Falade and his debut, Call Me, I’ll Hide the Body: after interest from 14 publishers and offers surpassing 2 million dollars, the agency representing the young writer decided to withdraw the crime novel due to “doubts about the manuscript’s genesis.” Falade, in addition to denying the use of AI, spoke of a “bias against Black authors.”
In March, instead, the publishing house Hachette decided to pull from the market the horror Shy Girl by Mia Ballard for a suspected use of AI. This case, which followed the publication of the novel that, as also reported by the BBC, had already sold almost two thousand copies in the British market. Although the use was always discussed as alleged, Ballard’s name remains marked: “This controversy has changed my life in many ways, my mental health is at a historic low and my name is ruined for something I did not even personally do,” the author revealed.
The Statement from Les Éditions du Boréal (Orélien’s Canadian Publisher)
But let us return to recount the Thélyson Orélien case: on September 24, the Canadian publisher of Orélien issued a statement to the press in which it commented on the affair, stating that “in none of these phases (linguistic revision and proofreading, ndr) Boréal has resorted to artificial intelligence.” On the other hand, regarding the work carried out before sending the manuscript, it is underscored that for the publisher “it is impossible to confirm or deny the accusations against the author” and, in these circumstances, it is essential not to draw hasty conclusions….” To add: “In the absence of concrete and tangible evidence, it is not up to us to decide whether he has the right to exist.”
Reflecting on the current debate, which goes beyond the single case and now involves many more key players, the Canadian publisher Boréal emphasizes that the publisher-author relationship must be based on trust: “It must not turn into a surveillance process.”
In this discussion, moreover, one should not forget the centrality of copyright in editorial dynamics…
The Response to the Plagiarism Allegation
With a text sent to the Agence France-Presse (AFP) in mid-October, Thélyson Orélien also expressed his position after the accusations and the countless comments. The Canadian-Haitian writer called the accusation of using artificial intelligence “a calumny and heresy,” explaining that he had used “tools for research,” but that “no sentence was created by a machine” (as also reported by Le Monde).
Regarding the plagiarism accusations, Orélien states that he “committed unacceptable errors,” apologizing for copying passages without citing sources.
Just a few weeks earlier, in a press release, his publisher had described these accusations as a “cause for concern,” because “they strike at values dear to us…”.
Paolo Giordano and the “Defense” Against AI
As mentioned earlier, even in Italy these topics are widely debated, and in this regard seems likely to open another branch of the discussion the reflection published by writer Paolo Giordano on October 13 in Corriere della Sera. The 2008 Strega Prize winner also commented on the Orélien affair and revealed his non-relationship with artificial intelligence.
If in a first pass Giordano described the Goncourt Academy’s choice as a “holy war – which we will lose – in defense of natural intelligence,” in a second remark he focused on the present: “[AI] Perhaps it will not wipe out novelists, but it has already changed something about our trade….”
Giordano, who disclosed that he is at work on a new novel, explained his method of saving the file: “I want to leave a clear trace, in case it proves necessary, of the progressive editing process, and also of my disorientation.”
“I Disconnected the Internet from the Computer I Write On”
But there is more. In response to many of the accusations and revelations concerning “the intrusion of social platforms into elections in various countries,” the impact on the health of social platforms and “the training of artificial intelligences” violating privacy and copyright, Paolo Giordano stated that he had “disconnected the computer I write on from the Internet.”
To protect himself from possible external visitors, Giordano chose to “go back,” avoiding cloud saves or sending drafts by email: “For the first time in my career, my agent, at the right moment, will not receive a message with an attachment. I will bring her the USB stick or printed sheets, in person, or I will send them by express courier. The same will be done by her with the publishing house.”
And so, between the fear of appearing paranoid and the explanation of the cost of his choices, the author of the long-selling The Solitude of Numbers concluded with a proposal. Not to “burn the machines (for the moment),” but to “use them as little as possible.”
The Debate Sparked by Giordano’s Reflection
Chiara Valerio, a writer, responded to Giordano with a piece for The Post: “I refuse to be verified and certified by a machine that has fed itself on my words and others’ for years. I refuse, in short, to accept that the chronology of my words is established by a machine.” The essayist goes on: “I know very well that someone might tell me tomorrow that what I wrote today, and the way I wrote it, is AI. But it is, on the contrary, the AI that comes from me, indeed from all of us.”
Giordano’s reply was echoed by D’Isa himself in a reflection published by theitalianreview.com titled “Dear Paolo, stop numbering the files,” which begins with: “(…) Giordano is completely free not to use artificial intelligence, to distrust it, to detest it, to write his novels on a clay tablet, if that makes him feel better. I believe no one, not even the most obsessed technophile, ever asked him to do otherwise. The trouble is that he asks something of us….” And later he writes: “(…) So I now offer Paolo and the crusaders a proposal: why not you keep writing as you like, stop numbering the files, and all together agree that everyone writes with whatever they want? You can reject these machines for ethical, aesthetic, or political reasons. So far so good. Just as long as those who do not use a tool do not claim the authority to decide how those who do use it should work….”
