Anthropic's rewritten Usage Policy takes effect on 12 November 2026. Anthropic announced it on 8 October. TechTimes reported on 9 November.
The new High-risk Physical Actions section applies when Claude drives equipment that could injure someone without a human approving each action. It sets three duties:
- A qualified person must be able to observe the equipment and stop it at any time.
- The equipment must stop or hold a safe state when that person intervenes or when the connection to Anthropic's services is lost.
- Limits such as speed, force, temperature or dose must be enforced by the equipment or by a controller independent of model output.
Consumer and household devices within their built-in safety limits are excluded. So are plans or commands that a qualified person reviews before execution, sensing without control, and laboratory automation with non-hazardous materials.
Terms such as "qualified individual" and "connection lost" are not defined. Customers who use Claude on AWS Bedrock accept the new terms by continuing to use it on or after 12 November. According to an AWS notice cited by TechTimes, model availability, pricing, API behaviour and model IDs do not change.
For readers of crypto-paid LLM services: this concerns Anthropic's own terms. It does not change prompt retention or training policies.