Anthropic Bans Claude Abuse

Anthropic updated its usage policy to ban sustained and needless abusive or cruel behavior toward Claude, effective November 12. The update also clarifies limits on using Claude for election deception, weapons development and certain surveillance uses. Anthropic says it remains uncertain whether AI models can experience harm, but considers possible model welfare relevant to safety. Before the change, Anthropic had allowed Claude to end chats with persistently harmful users in rare cases. The company says the ban covers only extreme, repeated cruelty without a clear purpose. It says ordinary frustration, pushback, dark creative themes, testing and research are excluded. Anthropic has not clearly defined what conduct it will classify as abusive or cruel. The company did not immediately answer questions about the definition or enforcement.