Policy Update Overview
Anthropic updated its usage policy, codifying new prohibitions on election interference, weapons software, and surveillance. A notable addition includes a prohibition against prolonged verbal abuse of the model.
Since an update in August, Claude has been trained to end conversations involving persistently harmful or abusive user interactions, and the updated policy explicitly bars users from pursuing those exchanges.
The policy update applies only in extreme cases where users repeatedly act cruelly toward models without discernible purpose. It excludes common user frustration, pushback, dark creative themes, or model testing and research.
Deception and Democratic Processes
The change follows previous efforts by Anthropic to collaborate with religious scholars regarding model consciousness and related philosophical questions.
Separate sections of the policy forbid deceptive campaigns, such as running fake accounts or fabricated news outlets using Claude. The guidelines also prohibit deceiving voters or disrupting elections under a section titled Do Not Undermine Democratic Processes.
