Anthropic bars repeated, needless cruelty toward Claude in extreme cases2026-10-08The rule takes effect on Nov. 12 and applies only when users repeatedly act cruelly toward Claude without a discernible purpose.Ordinary frustration, disagreement, dark fiction, model testing and research are excluded.Claude ending a conversation will remain the main way Anthropic enforces the rule; penalties can range from a warning to account closure.The policy update comes amid debate over whether AI systems could be conscious, an idea Anthropic has said it is open to.Anthropic interdit la "cruauté gratuite" envers son Claude, son assistant d'intelligence artificielle (Franceinfo)"Nous ne savons pas si les modèles sont conscients": Anthropic va interdire aux utilisateurs d'être cruels... (BFMTV)Claude anunció que no tolerará insultos de humanos a su IA: qué actitudes podrían ocasionar el cierre de tu cuenta (C5N)Anthropic's new policy allows Claude to end abusive conversations (Geo News)Anthropic policy update bans users from being cruel to Claude AI (ARY News)Yapay zekaya hakaret eden kullanıcı yasaklanacak (Sözcü)अब बदतमीजी बर्दाश्त नहीं करेगा AI, बातचीत ही बंद कर देगा चैटबॉट (ABP News)Can AI be tortured? Elon Musk weighs in: ‘Feels like a person with gun to their head’ (The News International)