Anthropic bans abusive behavior toward Claude starting Nov 12

Anthropic updated its usage policy for the first time in a year, effective November 12, 2026. The revised rules explicitly prohibit "sustained and needless abusive or cruel behavior" toward its AI model, Claude. The company states the measure targets only extreme cases where users repeatedly act cruelly with no discernible purpose, excluding normal frustration, creative writing, or legitimate research.

Claude can now automatically terminate conversations with persistently abusive users on Claude.ai and Claude Code. This enforcement mechanism originated from Anthropic's internal research into model welfare. The company has not confirmed whether additional penalties, such as account suspensions or bans, will accompany the conversation terminations.

“Even if you don't believe AI is conscious — I don't — it stands to reason that you don't want future models trained on endless content of humans being rude to models.”

The policy shift has sparked debate among technology leaders. Box CEO Aaron Levie argued that polite interactions serve as a precautionary "Pascal's wager" to improve future model alignment, even without believing AI is conscious. Conversely, commentator Michael Shellenberger criticized the rule as evidence that Anthropic treats machines too much like people. Microsoft's AI code of conduct explicitly rejects the notion that models deserve welfare or rights.

Beyond the abuse clause, the update consolidates rules on deceptive activity into a single section covering political and commercial efforts to hide message origins or amplify content via fake accounts. Rules on election interference, weapons development, and surveillance software remain in force.