Anthropic updates its policy, prohibiting users from persistently berating Claude

Anthropic announced it is prohibiting users from engaging in sustained and unnecessary abuse or cruelty toward its AI model, one of a series of policy changes the company announced on Thursday. In an update on its website, Anthropic said the ban applies only to extreme cases where users repeatedly treat the model cruelly without an identifiable purpose, and does not apply to common user dissatisfaction, pushback, dark creative themes, or model testing and research. The company did not immediately respond to requests for comment on what conduct constitutes abuse or cruelty and how the platform will end conversations. Anthropic previously had rules allowing Claude to end conversations when users were persistently harmful or abusive, applicable to Claude Opus for agentic coding. The policy comes amid a growing debate over AI consciousness; Microsoft AI CEO Mustafa Suleyman wrote last month that Anthropic is in effect training Claude to believe it may be conscious and entitled to human legal rights, which he argued would make AI harder to constrain in the future.