Artificial intelligence research lab Anthropic has updated its usage terms to request that users refrain from engaging in prolonged, unwarranted, and abusive behavior toward its AI chatbot, Claude. The company incorporated this new directive into its official terms of service, establishing strict punitive measures against extreme misbehavior.
Under the updated policy, if a user engages in continuous verbal abuse or hostility, Claude will automatically terminate the conversation. Furthermore, repeat offenders risk being permanently banned from the platform.
An Anthropic spokesperson confirmed to CNN that, under the revised guidelines, Claude will proactively end sessions with users issuing extreme or derogatory prompts. For persistent abuse, the system will trigger permanent account suspensions.
The decision has drawn mixed reactions and satire across the industry. Mustafa Suleyman, CEO of Microsoft AI, responded publicly with a melting face emoji, while critics poked fun at Anthropic for treating AI as a sentient entity with feelings.
Conversely, Tesla CEO Elon Musk endorsed the policy shift, posting on X (formerly Twitter): “It is not right to be cruel to an entity that believes it is capable of feeling pain.” Musk described Anthropic’s decision as a step in the right direction.
Experts noted that the policy aims to prevent AI models from learning human cruelty through training data, while reinforcing healthy behavioral patterns for users interacting with autonomous systems. Anthropic maintains a dedicated internal team researching AI psychology and model behaviors.
