Anthropic Bans Abusing Claude in New Usage Policy — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic updated its usage policy Thursday, codifying new prohibitions on election interference, weapons software, and surveillance
- A new provision explicitly bars users from pursuing prolonged verbal abuse of Claude, applying only in "extreme cases" where users repeatedly act cruelly with no discernible purpose
- Since an August update, Claude has already been trained to end conversations with "persistently harmful or abusive user interactions" — the new policy now formally bars users from restarting those conversations
- The policy explicitly carves out common user frustration, pushback, dark creative themes, and model testing or research from the abuse prohibition
- Separate rules ban using Claude to run fake accounts or fabricated news outlets, with election-related deception and disruption collected in a section titled "Do Not Undermine Democratic Processes"
- The policy shift follows Anthropic's collaborations with prominent religious scholars — in some cases seeking to convince them Claude could be conscious or have a soul
Why it matters: The update gives Anthropic a contractual basis to refuse service to users who persistently torment Claude, converting a model-level behavior (ending abusive chats) into a user-facing usage term. The narrow carve-outs for red-teaming, dark creative themes, and ordinary frustration reveal Anthropic is trying to police real cruelty without chilling legitimate research or creative use of the model.
Ask SkimNews
