Anthropic updates Usage Policy, effective Nov. 12

On Oct. 8, 2026 Anthropic published a Usage Policy refresh that takes effect November 12. The company added a ban on deceptive campaigns and fake-account infrastructure, clarified that weapons rules cover guidance software and arming drones, tightened non-consensual tracking and law-enforcement decision uses, and required a qualified operator who can stop hardware that can cause injury. It also prohibits sustained needless abuse of Claude, with conversation-ending on Claude.ai and Claude Code as the stated primary enforcement.