Anthropic updates its Usage Policy, adding rules on deceptive campaigns from November 12
3-line summary
- Anthropic published a new version of its Usage Policy for Claude, effective November 12.
- It consolidates rules against fake-account influence campaigns into one section and spells out its weapons and surveillance prohibitions more precisely.
- High-impact uses such as health and finance still require a qualified human reviewer and telling people that AI was used.
Anthropic published a new version of its Usage Policy for Claude on October 8. It takes effect on November 12, and most changes clarify existing rules. Because Claude now takes on longer, more independent work, the update adds examples showing how the rules apply to these new capabilities.
The biggest change is a new section, "Do Not Engage in Deceptive Campaigns or Artificial Activity." After seeing state media, government propaganda offices and commercial firms use Claude to run networks of fake accounts and fabricated news sites, Anthropic consolidated rules that had been spread across its elections, fraud, privacy and disinformation sections. Hiding who is behind a message, amplifying content with fake accounts — political or commercial — and building tools for such campaigns are all prohibited.
The weapons section now states that it covers the software and components that make weapons work and arming drones and other autonomous vehicles. The surveillance section prohibits tracking people without consent and using Claude to decide or recommend who to investigate, arrest or charge, while consented tracking such as fraud monitoring, content moderation, journalism and legal research remain allowed. The elections section is renamed "Do Not Undermine Democratic Processes," and a blanket ban on personalized vote and campaign targeting — which had blocked legitimate work like nonprofits writing voter information in other languages — has been removed.
For uses that affect someone's health, legal rights, finances or livelihood, Claude's recommendations must still be reviewed by a qualified person and the affected individual told that AI was used; the section has been rewritten to list which cases are covered. New requirements apply when Claude is connected to hardware that takes autonomous physical actions: an operator must be able to observe and stop the equipment, and it must hold a safe state if Claude disconnects. A new rule also prohibits sustained, needless cruelty toward the models in extreme cases, but not ordinary frustration or research testing.
Why it matters
As AI takes on more work directly, the rules on what it may be used for are becoming more concrete and example-driven. Companies building on Claude should recheck their use cases against the revised sections before the effective date.
Who it's for
- Developers
- Before November 12, check the newly listed high-risk use cases such as health, finance and legal; if yours qualifies, you need human review and AI disclosure.
- Marketers
- Campaigns that spread content through fake accounts or undisclosed posts are banned whether political or commercial — check this section if you run social content with AI.
Sources
- 2026 Usage Policy update — Anthropic
