Anthropic bans ‘abusive or cruel behavior’ toward Claude
The Verge Hayden Field ● Covered by 3 sources
Anthropic updated Claude’s rules and now bans abusive or cruel behavior toward it. It’s a sign the company is taking “model welfare” seriously, and using Claude for risky stuff is under tighter scrutiny too.
Based on reporting by The Verge, Hayden Field — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Anthropic has changed its usage policy for the first time in more than a year, and the new version is aimed at a wider set of misuse cases. The company says the update covers high-risk abuse tied to election interference, weapons development, surveillance, and health and financial uses.
The most eye-catching change is a ban on “sustained and needless abusive or cruel behavior” toward Claude. That’s a sharper line than most people expect from an AI policy, but it fits with Anthropic’s earlier work on what it calls “model welfare.”
Last August, the company said Claude would be allowed to end conversations with users who were “persistently harmful or abusive.” The new policy keeps that approach in place. Anthropic now says ending those chats remains the “primary enforcement mechanism.”
The company did not provide a comment on the update, according to The Verge. But the direction is clear: Anthropic is widening its safety net beyond obvious misuse and also drawing a boundary around how people can behave toward the model itself.
My take — AI-written commentary, not fact-checked reporting
Anthropic is doing the rare sensible thing here: treating AI misuse as both a technical and a social problem. The industry loves pretending abuse is just “user behavior” until it becomes convenient to call a model a partner, a therapist, or a teammate. Once a company opens that door, it can’t act shocked when the etiquette gets formalized.
Read more about this at: The Verge