Anthropic bans ‘abusive or cruel behavior’ towards Claude

Anthropic updated its usage policy to ban "sustained and needless abusive or cruel behavior" toward Claude, building on its model welfare research. The policy also consolidates restrictions on deceptive commercial and political campaigns, voter deception, election disruption, and AI-generated propaganda. It expands the weapons ban to cover software and components that make weapons work, and clarifies prohibitions on surveillance, including tracking people without consent or deciding who to investigate or arrest. Government contracts may be exempted if safeguards are adequate. Anthropic also requires a qualified operator to monitor and stop AI connected to hardware that could cause injury. The company continues to investigate questions about potential model consciousness and welfare.

(Source：The Verge)