Anthropic adds rule against abusing its AI models
The updated usage policy bars repeatedly directing unnecessary abuse or cruelty at Claude, but does not explicitly mention banning users.
Anthropic has updated its usage policy to prohibit repeatedly directing unnecessary abuse or cruelty toward its AI models, including Claude. The addition was revealed in a screenshot that Jane Manchun Wong shared on X.
According to the policy wording, the rule is aimed at extreme, deliberate abuse rather than everyday frustration, criticism, disagreements, or legitimate AI testing. Ordinary pushback from users and standard evaluation work fall outside its scope.
As the primary enforcement measure, Claude can end abusive conversations. The policy does not explicitly mention banning users, leaving the stated consequence at the level of terminating the exchange.
The change places limits on how users interact with Claude, drawing a line between critical prompting and repeated cruelty directed at the model.
Quick answers
What does Anthropic's updated usage policy prohibit?
It prohibits repeatedly directing unnecessary abuse or cruelty toward its AI models, including Claude.
What happens if a conversation turns abusive?
Claude can end the abusive conversation as the primary enforcement measure. The policy does not explicitly mention banning users.
Does the rule cover criticism or AI testing?
No. It targets extreme, deliberate abuse, not everyday frustration, criticism, disagreements, or legitimate AI testing.