Un auteur continue d'utiliser l'IA malgré des citations synthétiques dans son livre

Steven Rosenbaum a admis que des outils d'IA avaient introduit quelques citations synthétiques dans son nouveau livre sur la vérité et la réalité. Ces erreurs ont été révélées lors d'une récente enquête du New York Times. Rosenbaum a déclaré être désormais plus prudent, mais prévoit de continuer à s'appuyer sur cette technologie.

Le livre de Rosenbaum, The Future of Truth: How AI Reshapes Reality, traite de la manière dont l'intelligence artificielle déforme les faits. Sur ses 285 références, six ont été signalées comme problématiques, dont trois citations synthétiques sans source claire. Une citation attribuée à la journaliste technologique Kara Swisher a été démentie par celle-ci, tandis qu'une autre, liée à la professeure Lisa Feldman Barrett, a été qualifiée d'inexacte et inexistante dans ses travaux.

Articles connexes

Illustration of lawyers in court using AI for fake citations in a Meta Facebook lawsuit case.
Image générée par IA

Lawyers face sanctions for AI fake citations in Facebook lawsuit

Rapporté par l'IA Image générée par IA

A US appeals court has warned that lawyers may face sanctions after submitting an appeal filled with fictitious quotations generated by artificial intelligence. The case involved an attempt to force Meta to remove a critical post from a dating safety group on Facebook.

Nobel laureate Olga Tokarczuk has responded to controversy over her reputed use of AI in writing her latest novel.

Rapporté par l'IA

New research indicates that conversational AI can strengthen false beliefs and distorted thinking in users. The findings come from a study examining how chatbots interact with people experiencing delusions or conspiracy theories. Researchers highlight risks especially for those who are isolated or vulnerable.

Over 10,000 writers have collaborated on an empty book to protest the unauthorized use of their work in training AI models. The initiative, led by composer Ed Newton-Rex, aims to influence UK copyright law discussions. Copies will be distributed at the London Book Fair to raise awareness.

Rapporté par l'IA

Leading AI coding assistants fail one in four tasks, according to a TechRadar analysis. The report points to serious gaps between hype and actual performance reliability, especially in structured output tasks. AI tools are far from flawless in these critical areas.

South Africa's Communications Minister Solly Malatsi has withdrawn the draft National Artificial Intelligence Policy following revelations of fictitious sources in its references, likely generated by AI tools. The errors impacted three of the policy's six pillars, leading to internal probes and commitments to accountability. Malatsi described the lapse as a key reason for needing stronger human oversight in AI use.

Rapporté par l'IA

Three of five regional winners of the 2026 Commonwealth Short Story Prize are under suspicion for using generative AI in their entries. The accusations have shifted attention from celebration to controversy within the literary community.

 

 

 

Ce site utilise des cookies

Nous utilisons des cookies pour l'analyse afin d'améliorer notre site. Lisez notre politique de confidentialité pour plus d'informations.
Refuser