Anthropic's new usage policy takes effect on 12 November
On 8 October, Anthropic published a new version of its Usage Policy, the rules for what people and companies may and may not do with Claude. The updated policy takes effect on 12 November 2026. Anthropic says most of the changes clarify existing rules, with new examples for the longer and more independent work Claude now does.
If your business builds anything on Claude, this policy is one of the sets of rules your use has to follow. A few changes are worth reading closely before the deadline.
Fake accounts and deceptive campaigns
Anthropic says it has seen state media outlets, government propaganda offices and commercial firms use Claude to run networks of fake accounts and fabricated news sites. The rules against this were spread across several sections. They now sit in one section, titled Do Not Engage in Deceptive Campaigns or Artificial Activity, which covers political and commercial cases alike, including hiding who is behind a message, amplifying content with fake accounts and building the tools for these campaigns.
The elections section is now narrower, focused on deceiving voters or disrupting elections. Anthropic removed the blanket ban on personalized vote and campaign targeting, saying it blocked legitimate civic work such as nonprofits writing voter information in other languages.
Health, money and legal rights
The requirements for high-risk uses did not change, but Anthropic rewrote the section to make them easier to apply. When Claude is used in ways that can affect someone's health, legal rights, finances, livelihood or access to essential services, there must be a qualified person who can review and change Claude's recommendation, and the person affected must be told that AI was used. The section now lists which kinds of recommendations are covered and which are not.
There are also new rules for Claude connected to hardware that acts in the physical world and could injure someone. A qualified operator must be able to watch the equipment and stop it, and the equipment must be able to hold a safe state if Claude disconnects.
Other changes
The weapons section now says explicitly that it covers the software and parts that make weapons work, and actions like arming drones. The surveillance section bans tracking people without consent and using Claude to decide or recommend who to investigate, arrest or charge, while keeping consented uses such as fraud monitoring. A new rule targets sustained, pointless abuse of the models, which Anthropic says applies only in extreme cases. The Supported Regions Policy now states that companies majority-owned or controlled from unsupported regions cannot use Claude.
What to do before 12 November
If your product uses Claude to make or suggest decisions about people's money, health, jobs or legal matters, read the rewritten high-risk section and check two things: that a qualified person can change the result, and that the affected person is told AI was involved. If you run marketing or social media tools on Claude, read the new deceptive campaigns section with your team.
What we think
Most companies will not notice this update, and that is fine. The high-risk section is the one that matters for ordinary business use, and it helps that Anthropic now writes down which recommendations are covered, so customers do not have to guess.
Sources