2026 Usage Policy update

Summary

Anthropic has published an updated Usage Policy effective November 12, 2026, reflecting Claude’s evolving capabilities and observed misuse patterns in influence operations, weapons development, and surveillance. The update consolidates rules against deceptive activity into a new section titled Do Not Engage in Deceptive Campaigns or Artificial Activity. The elections section is refocused to specifically disallow deceiving voters or disrupting elections, removing a blanket prohibition on personalized targeting that hindered legitimate civic work. Weapons prohibitions now explicitly include software and components for weapons, as well as arming autonomous vehicles, and the surveillance section more precisely prohibits tracking without consent and using Claude for law enforcement decisions. The policy also clarifies high-risk use case requirements, adds controls for autonomous physical actions requiring a qualified operator and a safe state, and prohibits sustained abusive behavior toward models.

Takeaways

1. 1

   The new Usage Policy consolidates rules against deceptive activity into a single section prohibiting deceptive campaigns or artificial activity, covering influence operations and fake accounts.

2. 2

   The elections section is refocused to prohibit deceiving voters or disrupting elections, removing a blanket prohibition on personalized targeting that previously hindered legitimate civic work like multilingual voter information.

3. 3

   Weapons prohibitions now explicitly include software and components for weapons, as well as arming autonomous vehicles, and the surveillance section more precisely prohibits tracking without consent and using Claude for law enforcement decisions.

FAQ