BUSINESSCompanies · · 2 min read · Source: Anthropic
Anthropic: new Claude usage policy takes effect November 12
Anthropic has revised its usage policy for Claude, effective November 12, 2026. Changes include a new section on deceptive campaigns, tighter surveillance rules and a ban on sustained, needless abuse of the model.
Key facts
- Anthropic's revised Claude usage policy takes effect on November 12, 2026.
- A new section prohibits deceptive campaigns using fake accounts or artificial amplification, both political and commercial.
- A new rule bans sustained, needless abuse of the model; ordinary frustration, criticism and research are excluded.
What is changing
The policy governs how Claude may be used in the app, in Claude Code and via the API. Anthropic describes the update mainly as a clarification. The key points:
- Deceptive campaigns: A new section brings together rules on fake accounts, fabricated news sites and artificial amplification. It explicitly covers political and commercial activity, which includes marketing.
- Elections: The section now targets voter deception and election disruption. The blanket ban on personalized campaign targeting has been removed.
- Surveillance: Tracking people without consent is prohibited, whether in real time or from data already collected. Claude may not decide or recommend whom to investigate.
- High-risk uses: Where Claude affects people's health, rights, finances or livelihood, a qualified human must still review the output and the people affected must be told AI was involved.
- Hardware: If Claude controls equipment that acts on its own, a human must be able to watch and stop it.
The rule everyone is talking about
A new prohibition covers sustained, needless and cruel behavior toward the model. Anthropic says it targets only extreme, repeated cases with no discernible purpose. Ordinary frustration, pushback, dark creative themes and testing or research are excluded. The main enforcement tool remains Claude's ability to end such conversations. CNN reported on November 1 that bans are possible in extreme cases.
Regions and companies
Anthropic also clarifies that companies based in unsupported regions, or majority-owned by entities there, may not use Claude. For most US and European businesses nothing changes, but international group structures are worth checking against the list.
What this means for you
- Fake reviews at scale, sock-puppet accounts and bought reach are explicitly off-limits with Claude, including in commercial marketing.
- If you use Claude for decisions about loans, job applications or similar, you need a qualified human reviewer and must inform the people affected.
- Everyday work with Claude barely changes: being annoyed at a bad answer is explicitly allowed.
How to use this
Write half a page of AI rules for your team: what you use Claude for, where a human checks and when you tell customers AI was involved. For decisions about people, such as hiring or lending, the policy requires human review.
Luigi's take
The headlines are about the abuse ban. For businesses, the section on deceptive campaigns matters more. Naming commercial activity explicitly sends a clear signal to anyone planning to use AI for fake profiles or artificial reach. In 30 years of marketing I have seen plenty of shortcuts come and go. Bought reach was never one that paid off in the long run, and skeptical audiences spot it quickly.
My advice: if your team uses Claude day to day, use the new policy as a prompt to write a short internal rule. What do we use AI for, where does a human check, when do we tell customers AI was involved? Half a page is enough. It keeps you on the right side of Anthropic and, more importantly, of your customers' trust.
Source
Anthropic: 2026 Usage Policy update
More sources
- CNN – Bericht vom 1. November 2026 zur neuen Regel gegen Missbrauch von Claude
This article was researched and written with AI assistance. Responsible: Luigi Lugmayr.
