Anthropic revises usage policy to ban repeated abuse of Claude models

Anthropic will prohibit users from repeatedly abusing its Claude models without meaningful purpose, with the new usage policy taking effect on November 12, 2026. The restriction adds an unusual dimension to AI safety rules, placing abuse of chatbots alongside prohibitions on harassment and other harmful conduct. The San Francisco-based AI company said the change targets extreme cases, not ordinary disappointment with chatbot responses. Users can still challenge Claude's answers, test its limits, and explore dark themes in creative work.

Ban on meaningless abusive behavior

Anthropic defined the boundaries of abuse, with the updated policy prohibiting "persistent and unnecessary abuse or cruel behavior" toward its models. The company said the rule applies only when users repeatedly exhibit cruelty without apparent purpose. Anthropic previously discussed model welfare in August 2025, studying whether Claude should end conversations with persistently harmful or abusive users. Conversation termination remains the primary enforcement mechanism under the revised policy. Anthropic has not clarified whether repeat violations would trigger additional penalties such as account suspension.

The broader policy also prohibits behavior causing psychological harm to people. Its restrictions cover promotion of suicide and self-harm, encouragement of eating disorders, body shaming, non-consensual intimate imagery, harassment, and animal abuse. One provision also targets products designed to cause emotional harm through deceptive techniques. The restriction extends the policy beyond individual conversations to encompass services built around harmful interactions. In addition to targeting abuse, Anthropic has integrated restrictions on deceptive commercial and political activity; the revised rules prohibit the use of fake accounts or posts to conceal who sponsors a message or artificially amplify its reach.

Expanded psychological harm restrictions

Election-related restrictions cover voter deception, candidate impersonation, and attempts to suppress turnout. The company also prohibits misinformation about candidates and voting procedures. Weapons restrictions now explicitly cover software and components that enable weapons to function. The policy also addresses activities such as armed drones and other autonomous vehicles. Following concerns about AI-assisted tracking, Anthropic has clarified its surveillance restrictions. Users may not use Claude to track non-consenting individuals, including through previously collected data. The policy also bars Claude from recommending whom authorities should investigate, arrest, or charge, and building or improving surveillance tools is also restricted.

Another provision addresses AI models connected to physical devices capable of causing harm. Anthropic requires qualified operators to observe such devices and halt their operation when necessary. The wording does not explicitly require the operator to be human. However, the requirement introduces direct oversight for AI systems controlling potentially dangerous machinery. Anthropic allows exceptions for certain government contracts when contractual restrictions and safeguards sufficiently reduce potential harm. The company has previously worked with the U.S. military.

Debate over AI consciousness

The updated policy also reflects the ongoing debate over AI consciousness. Anthropic researchers have investigated whether increasingly powerful models may possess morally relevant internal experiences. Chief Executive Dario Amodei acknowledged this uncertainty in a February 2026 podcast, stating, "We don't know if these models have consciousness." Microsoft's AI takes a more skeptical stance on model welfare; its code of conduct rejects legal personhood and rights for AI models, even as it acknowledges that consciousness research remains unresolved.

Anthropic's November deadline will bring its revised restrictions into force, with the policy distinguishing ordinary chatbot disappointment from repeated cruelty while extending safeguards across digital systems and physical machinery.

Impact of the new policy on AI use

Anthropic's new policy marks a strict regulatory framework for AI use, particularly in preventing abuse and protecting users. As AI technology advances, attention to its potential risks is growing; these policies target not only individual user behavior but also extend to entire interaction services. This reflects the industry's emphasis on AI ethics, especially in preventing psychological harm to humans and abusive conduct. As the exploration of AI consciousness continues, companies must consider these complex ethical issues when formulating policies.