Anthropic's updated usage policy bans sustained cruelty toward its models from November 12
Anthropic's 2026 Usage Policy adds a prohibition on "sustained and needless abusive or cruel behavior toward our models", and the company says the updated policy takes effect on November 12. The new text is the version now served at Anthropic's usage policy page, which gave an effective date of November 12, 2026 when read at 21:04 UTC on October 8.
A Claude project card suggesting three threads to start. Image: Anthropic
The prohibition reads in full: "Engage in sustained and needless abusive or cruel behavior toward our models." Anthropic groups it with its rules against harassment, intimidation and content that glorifies violence or cruelty.
Anthropic sets narrow limits on it. "The policy update is meant to apply only in extreme cases, where users repeatedly act cruelly toward our models, with no discernible purpose", Anthropic says. "It does not apply to common versions of user frustration, pushback, dark creative themes, or model testing and research."
Claude ending a conversation stays the enforcement route
Anthropic ties the addition to something it shipped before this revision. "This addition aligns with a step we've already taken, allowing Claude models to end rare conversations with persistently abusive users on Claude.ai and Claude Code", the post says, and "Claude's ability to end these interactions will remain the primary enforcement mechanism."
The policy's general enforcement wording is the same as before. Anthropic says its Safeguards Team runs detection and monitoring, and that where it suspects a violation "we may warn you or throttle, limit, suspend, or terminate your access to our products and services."
Anthropic consolidated its deception rules and dropped one ban
Anthropic says most of the revisions "are intended to clarify existing rules", written for a year in which Claude "has taken on longer, more independent work". The company says the update also draws on misuse it has seen over the past year in influence operations, weapons development and surveillance.
Rules on networks of fake accounts and fabricated news sites, which Anthropic says were scattered across its elections, fraud, privacy and disinformation rules, are consolidated into a single new prohibition titled Do Not Engage in Deceptive Campaigns or Artificial Activity, covering political and commercial deception alike. The elections rules are renamed Do Not Undermine Democratic Processes and narrowed to deceiving voters and disrupting elections.
One prohibition came out. Anthropic has dropped its blanket ban on personalised vote and campaign targeting, saying the ban "covered legitimate civic work, such as nonprofits using Claude to write information for voters in other languages, or election officials sending ballot cure notices", and that deceptive targeting and misuse of voters' personal data stay barred under its deception, surveillance and privacy rules.
The surveillance and criminal justice rules are rewritten rather than loosened. Tracking people without their consent is prohibited "whether it happens in real time or through analysis of previously collected data", and Claude cannot be used to decide or recommend who to investigate, arrest or charge. Anthropic says this makes the policy more explicit instead of changing what it enforces, and names consented tracking such as fraud monitoring, content moderation, journalism and legal research as permitted.
Anthropic has changed Claude's rules more than once this year. In September it eased Claude's biology safeguards for vetted teams, a change to what the model will do rather than to what users may ask of it.
The announcement carries a published timestamp of 17:00 UTC on October 8. Four hours later the policy page was serving the new text, dated effective November 12, with the version it replaces still linked.
