Anthropic updates its usage policy, prohibiting model abuse and election interference

Anthropic updated its usage policy on Thursday, adding prohibitions on election interference, weapons software, and surveillance, and explicitly prohibiting users from engaging in long-term verbal abuse of the model. Since the August update, Claude had already been trained to end conversations when faced with sustained harmful or abusive interactions; the new policy directly prohibits users from initiating such conversations. Anthropic stated that this clause applies only to extreme cases where users repeatedly treat the model maliciously without an identifiable purpose, and does not apply to common user frustration, pushback, dark creative themes, or model testing and research. Other sections of the policy also prohibit using Claude for deceptive activities such as running fake accounts or fabricating news organizations, and include a section titled "Don't undermine the democratic process" prohibiting deceiving voters or interfering with elections.