Anthropic bans 'cruel' behaviour against its Claude AI

Already last year, the company said it was giving Claude the ability to end conversations in rare, extreme cases of "persistently harmful or abusive user interactions”