Anthropic Bans Abuse of Claude

Anthropic has officially banned abusive behavior towards the language model Claude. The updated terms of service now include a clause prohibiting systematic and purposeless cruel or offensive treatment of AI.

Users can no longer bully the model through pointless matrix multiplication. If a user relentlessly torments Claude without purpose, the system reserves the right to terminate the conversation and cease responding within the current session.

What Remains Permitted

The new rules do not ban casual swearing in chat. Harsh fictional scenarios and standard model testing remain acceptable. The ban targets only purposeless cruelty, not productive work with AI.

Anthropic explains that the company is still unsure whether models can experience emotions. Developers aim to determine if artificial intelligence welfare deserves creator attention. The new rules take effect on November 12, 2026.