Overview
- Anthropic announced the usage policy update on Thursday, October 8, 2026, and said the revised rules will take effect on November 12, 2026 to give developers and partners time to adapt.
- The policy adds a narrow ban on “sustained and needless abusive or cruel behavior” toward Claude and codifies an August capability that lets Claude end persistently abusive conversations.
- Enforcement relies primarily on terminating conversations with offending users and the company did not commit to broader penalties such as account bans or public enforcement processes.
- The refresh tightens prohibitions on deceptive influence by banning voter deception and impersonation, expands weapons limits to bar software and components that enable weapons or armed drones, and forbids non-consensual tracking and use of Claude for law‑enforcement charging decisions.
- Anthropic says the rules can be modified for certain government contracts if contractual safeguards are deemed adequate, a choice that reflects tensions between safety, commercial partnerships and calls for clearer industry-wide safeguards after documented misuse attempts.