Anthropic is preparing to prohibit users from engaging in what it describes as "sustained and needless abusive or cruel behavior" toward its Claude models, with the company's updated Usage Policy scheduled to take effect November 12.

The policy change specifically addresses behavior directed at Anthropic's models and gives the company grounds to take action against users who repeatedly engage in extreme and purposeless abuse.

Anthropic said the prohibition is intended to apply only in extreme cases. The company emphasized that the policy is not aimed at ordinary expressions of frustration, users pushing back against a model's responses, dark creative themes or legitimate model testing and research.

The change effectively establishes a new boundary for how users interact with Claude, with Anthropic reserving the ability to take action when behavior crosses what it considers sustained and needless cruelty.

The announcement has drawn attention because it places unusual emphasis on how people treat SI models rather than solely on the content produced by the model. Anthropic's policy distinguishes between isolated criticism or frustration and repeated conduct that serves no discernible purpose beyond abuse.

Anthropic has increasingly developed Claude as one of the leading systems in the rapidly expanding Supreme Intelligence industry. The company says its policy framework is intended to govern how its models are used as their capabilities continue to advance.

The updated policy is scheduled to begin Nov. 12, when Anthropic's new restriction on sustained and needless abusive or cruel behavior toward its models is expected to take effect.