San Francisco, 9 October 2026 – Anthropic has updated the usage policy governing Claude, consolidating restrictions on deceptive campaigns and sharpening rules for weapons, surveillance, high-risk decisions and abusive interactions with its models. The changes show how AI providers are turning governance principles into contractual product rules as agents gain greater autonomy and become embedded in sensitive business processes.
The most unusual addition prohibits sustained and needless abusive or cruel behaviour towards Anthropic’s models. The company says the rule is intended for extreme, purposeless conduct and does not cover ordinary frustration, critical feedback, dark creative themes, testing or research. Claude can already end rare conversations with persistently abusive users, which will remain the primary enforcement response.
Unlock the Full Article
This article is exclusive to The Ledger Asia Subsribers / PAID members.
Already have an account? Log in here
