Research note
Research note: Anthropic’s new usage policy bans sustained abuse of Claude from 12 November
How we checked Anthropic’s usage policy changes against the policy text and Anthropic’s announcement.
What we set out to check
The effective date, the new abuse rule and its limits, and the changes to election, weapons and surveillance rules.
What we found
- The policy text says “Effective November 12, 2026”.
- It contains the line “sustained and needless abusive or cruel behavior toward our models” and a section titled “Do Not Undermine Democratic Processes”.
- Anthropic’s announcement explains that the abuse rule applies only in extreme cases and does not cover frustration, pushback, dark creative themes or research.
Labels we chose and why
- All policy facts are Confirmed: they come directly from Anthropic’s policy and announcement.
Conflict of interest
Anawords uses Claude, made by Anthropic, as a research tool. We stated this in the story. Anthropic had no role in the story or this note.
What we could not verify
- Whether breaking the abuse rule can lead to account suspension.
What we searched
- AI news this week October 2026
- Anthropic usage policy update November 12 abusive behavior democratic processes
Every source we read
| Source | Used? | Why |
|---|---|---|
| Anthropic: 2026 usage policy update | USED | Primary source explaining each change. Used for scope, exclusions and enforcement. |
| Anthropic: Usage Policy | USED | The policy text. Used to confirm the effective date and exact wording. |
| MediaNama: Anthropic blocks abusive behaviour towards Claude | USED | Used for context and for the open questions it put to Anthropic. |
| Quartz: Anthropic bans abusive behavior toward Claude | NOT USED | Less detail than the primary sources, so we did not need it. |