European publishing professionals are becoming more willing to discuss the use of artificial intelligence, even as concerns remain over authorship, detection and the risk of deception. Those divisions were on display at the second European Publishers’ Colloquium, held as part of the Formentor Literary Conversations in Pollença, Mallorca, with 28 editors, literary agents and scouts taking part.
Paloma Sánchez van Dijck, an editor at Dutch publisher Meulenhoff Boekerij, said the company reviews its AI policies every three weeks. It had previously ruled out using the technology for any material that could be written, but now uses it to prepare marketing copy. She said younger staff members had brought interesting ideas to the debate.
Participants appeared more open to discussing AI than at last year’s colloquium, where resistance was stronger. Transparency—making clear whether AI was used and for what purpose—surfaced repeatedly as a guiding principle.
Gustavo Guerrero, Gallimard’s editor for Spanish, Portuguese and Ibero-American publishing, argued that AI is here to stay and that the industry must learn how to deal with it. He cited the case of Canadian author Thélyson Orélien, whose book was withdrawn from consideration for the Goncourt prize after its use of AI was brought to light. Guerrero said technology had exposed the issue, while warning that detection software such as Pangram could eventually be overtaken or circumvented.
Gianluca Foglia, managing director of Italy’s Feltrinelli group, likewise predicted that AI use would become commonplace. In his view, the important question is not who uses it, but for what purpose.
Other speakers urged publishers to draw firm boundaries. Joaquín Palau of Spanish imprint Arpa said the industry must be prepared to reject exceptional manuscripts if they were produced with AI. He also argued that the challenge now extends beyond plagiarism to deliberate deception.
Some participants saw potential benefits. Literary agent Michael Gaeb described AI as an opportunity for serious literature, arguing that while machines can write well, they lack the quality that distinguishes great literature. Piero Salabé, a senior fiction editor at Germany’s Hanser, suggested AI could help readers distinguish art from formula-driven work.
The discussion also addressed disputed cases and the limits of detection tools. Peruvian author David Silva Rojas’s award-winning novel Donde el agua trabaja was alleged to be 99% AI-generated, a claim he denies. Juliette Ponce of Éditions Dalva noted the paradox of people having to prove they are human to a machine, while literary scout Camilla Dubini pointed to differences between uses such as translation and generating content. Juan Cerezo, editorial director at Tusquets, proposed that the industry agree on a shared forensic tool to identify AI use.
Basilio Baltasar, a leading figure behind the Formentor conversations, closed the first session with a stark warning. He said he was not a pessimist but an apocalyptic, describing AI as “the embodiment of the devil.”
Sources: El País Cultura.