Anthropic has updated its usage policy to allow its artificial intelligence assistant, Claude, to end conversations in cases of sustained and unnecessary abusive behaviour. The company said the restriction would apply only in extreme cases involving repeated abuse. The updated rules add such conduct to a list of prohibited activities that also includes bullying, promoting self-harm and creating intimate images without consent. Anthropic clarified that ordinary frustration, criticism, dark creative themes and legitimate model testing would not fall under the restriction. The announcement has renewed debate over how people should interact with artificial intelligence systems and whether treating chatbots as if they have human feelings is appropriate. Critics argue that describing such behaviour as cruelty could encourage people to view software as human, while supporters believe the policy may encourage more respectful communication. Anthropic said the changes followed its annual review of usage policies and also addressed risks such as deceptive campaigns and attempts to mislead voters or disrupt elections. The debate highlights broader questions about responsible artificial intelligence use, online behaviour and the boundaries companies should establish for interactions with their tools.
Read more here
