On Thursday, San Francisco‑based AI lab Anthropic updated its usage policy to ban sustained and needless abusive or cruel behavior toward its Claude AI system.
The policy, which does not explicitly reference ‘model welfare’, still signals the company’s openness to protecting AI systems from harm.
Claude will retain its ability to end conversations when faced with persistent harmful or abusive user interactions, according to the policy.
"We don't know if the models are conscious…But we’re open to the idea that it could be," said CEO Dario Amodei in a February interview with The New York Times.
Industry voices remain divided. Microsoft AI chief Mustafa Suleyman dismissed the notion of machine consciousness, while Pope Leo XIV, in a sermon at St. Peter’s Basilica, argued that AI lacks a soul.
Users who violate the new terms will see Claude terminate the dialogue automatically, with no further interaction allowed, according to the updated enforcement mechanisms.