Anthropic updated its usage policy on Thursday, codifying new prohibitions on election interference, weapons software, and surveillance. But perhaps the most notable change is a new ban on long-term abuse of models, which is now explicitly prohibited.
Since the August update, Claude has been trained to end conversations that involve “persistently harmful or abusive user interactions,” but the new policy explicitly prohibits users from continuing such conversations.
“This policy update is intended to apply only in extreme cases where users repeatedly engage in cruel behavior towards our models for no apparent purpose,” the post explains. “Does not apply to user complaints, backlash, dark creative themes, general version of model testing and research.”
This change was made after Antropic sought collaboration with prominent religious scholars. In some cases, they were trying to convince them that Claude was conscious, or that he might have a soul.
Another section of our usage policy prohibits users from participating in a wide range of deceptive campaigns, such as using Claude to run fake accounts or fabricated news outlets. The policy also includes rules against voter fraud and election interference, summarized in a section titled “Do not undermine the democratic process.”
