Anthropic Updates Usage Policy with Ban on Abusive Behavior Toward Claude
Anthropic has updated its usage policy for the first time in over a year, introducing new restrictions to address high-risk misuse scenarios. One of the most notable changes is a ban on 'sustained and needless abusive or cruel behavior' toward its AI model, Claude. The company previously allowed Claude to terminate conversations with harmful users as part of research into 'model welfare.' This update reinforces that approach but clarifies it applies only in extreme cases. The policy also expands existing restrictions by banning deceptive political campaigns, voter deception during elections, and the use of Claude for weapons development or surveillance. The company has observed attempts to leverage its AI for creating guidance software for autonomous weapons systems and tracking individuals without consent. These new rules aim to prevent misuse while allowing flexibility in contracts with government clients under specific safeguards. Additionally, Anthropic introduced a rule requiring human oversight when its models are connected to hardware capable of causing injury. This aligns with broader industry efforts to ensure responsible AI deployment.
