Three former OpenAI researchers sound the alarm over AI with opaque reasoning2026-10-08Jasmine Wang, Tomek Korbak and Mikita Balesni, who were fired by OpenAI, are urging the company not to develop AI whose reasoning would be unreadable to the humans tasked with overseeing it.The three researchers say their dismissal, which OpenAI says was over leaks of sensitive information, risks discouraging employees from voicing their concerns.An OpenAI research executive said he agreed that it should be possible to verify, step by step, how models make their decisions.Intelligence artificielle : trois chercheurs licenciés par OpenAI l’appellent à ne pas créer d’IA au raisonnement illisible (Sud Ouest)"Nous ne savons pas encore comment développer et déployer en toute sécurité des modèles que nous ne pouvons... (BFMTV)OpenAI rejects researchers' claim they were fired over safety warnings, says they committed 'breach of trust' (The New York Post)Trois employés limogés par OpenAI l’accusent de faire taire ses spécialistes de la sécurité, l’entreprise dément (Le Monde)برطرف 3 محققین کے الزمات پر اوپن اے آئی کی وضاحت (Daily Jang)