Anthropic bans sustained, needless cruelty toward its models

Anthropic’s updated usage terms prohibit users from engaging in “sustained and needless abusive or cruel behavior toward our models.” The company announced the change on 8 October, and the terms took effect on 12 November.

Anthropic says the restriction is meant for extreme cases in which users repeatedly act cruelly toward its models for no discernible purpose. It does not cover ordinary frustration or pushback, dark creative themes, or model testing and research.

How the rule will be enforced

Before the update, Anthropic had already allowed Claude models to end rare conversations with persistently abusive users on Claude.ai and Claude Code. The company says Claude’s ability to end such interactions will remain the primary enforcement mechanism.

Anthropic has discretion over how it applies its terms, The Register said. The outlet also said that attempts to bypass Claude’s guardrails—restrictions on what the system will do—violate a separate platform-abuse prohibition, even though they have a discernible purpose.

The question of model welfare

Anthropic has said it remains deeply uncertain about model welfare and that there is no scientific consensus on whether current or future superintelligence (SI) systems could be conscious or have experiences deserving consideration. Its research program is exploring when, or whether, model welfare deserves moral consideration the potential importance of model preferences and signs of distress, and what practical, low-cost measures might be appropriate.