Anthropic updates usage policy to restrict model abuse and election interference

The company clarified that these restrictions target extreme, purposeful harassment rather than typical user frustration or creative experimentation. Since August, Claude has been equipped to terminate conversations involving persistently harmful input, but the new mandate codifies this as a violation of terms. This shift coincides with Anthropic’s recent efforts to engage religious scholars regarding the nature of AI consciousness and potential personhood.

Beyond interpersonal conduct, the policy establishes a clear stance on digital integrity. Users are now prohibited from leveraging the model for deceptive campaigns, such as managing networks of fake accounts or generating fabricated news. A dedicated section, "Do Not Undermine Democratic Processes," explicitly forbids any activity intended to manipulate voters or disrupt electoral systems, signaling a proactive approach to mitigating the misuse of generative AI in sensitive political environments.