Key facts
- Anthropic updated its usage policy on Thursday, October 8, 2026.
- The new policy prohibits election interference, weapons software, and surveillance.
- Users are now expressly forbidden from prolonged verbal abuse of Anthropic's models.
- The policy targets extreme cases of users repeatedly acting cruelly toward models.
- The policy does not apply to common user frustration, pushback, dark creative themes, or model testing.
- The policy includes rules against deceptive campaigns, such as using models for fake accounts or fabricated news.
AI company Anthropic has updated its usage policy, introducing new prohibitions against election interference, the development of weapons software, and surveillance. The policy also explicitly forbids prolonged verbal abuse of its AI models.
The policy update is intended for extreme cases where users repeatedly engage in cruel interactions with the models without a clear purpose, and does not cover typical user frustration, testing, or research. This change follows Anthropic's engagement with religious scholars regarding the potential consciousness or soul of its models.
Additionally, the updated policy prohibits users from conducting deceptive campaigns, such as operating fake accounts or generating fabricated news. It also includes specific rules against deceiving voters or disrupting democratic processes.