The development
What happened
Anthropic's updated rules govern use of Claude AI models, including deceptive campaigns, high-risk recommendations and AI-connected hardware that can act physically. Announced on 8 October, they take effect on 12 November. Many changes clarify existing rules; the announcement is not a new model launch.
Key facts
Announced on 8 October, the updated policy takes effect on 12 November 2026.
The updated election section focuses on deception and disruption; the blanket ban on personalized vote or campaign targeting is removed, while deception, privacy and surveillance restrictions remain.
High-risk recommendations retain requirements for qualified human review and notification to affected people.
For autonomous hardware capable of causing injury, a qualified operator must observe and be able to stop it, and equipment must hold a safe state if Claude disconnects.
A consolidated deceptive-campaign section covers political and commercial fake accounts, hidden message origins and artificial amplification.
The context
Why it matters
Teams deploying Claude AI models need to connect provider rules to actual model behavior: human review of consequential recommendations and observable, stoppable physical actions. Narrower election wording remains bounded by deception and privacy restrictions. These service rules do not by themselves establish legal compliance.