Anthropic adds a ban on abusive or cruel behavior toward Claude

The updated policy also tightens rules on weapons, surveillance and autonomous hardware, changing how Claude can be used in higher-risk settings.

Anthropic has updated Claude’s usage policy for the first time in over a year, adding tighter rules around abusive use, propaganda, surveillance, weapons, and autonomous hardware. The most unusual change bars sustained and needless cruelty toward Claude, while Anthropic says conversation-ending remains its main enforcement tool and that the rule is aimed only at extreme cases. The policy also consolidates bans on deceptive political or commercial campaigns, extends the weapons restriction to software and components that make weapons work, and clarifies that Claude cannot be used to track people without consent or to support surveillance tools. For systems tied to hardware that can act on its own and could injure someone, Anthropic now requires a qualified operator who can monitor the equipment and stop it if needed. The update follows Anthropic’s earlier model-welfare work, including last August’s move to let Claude end chats with persistently harmful users.