On October 8, 2026, Anthropic published a revised Usage Policy for its Claude models. The company says the update clarifies existing prohibitions and requirements in light of new usage patterns and customer feedback observed over the past year. The changes focus on deceptive campaigns, weapons-related uses, surveillance, high-risk applications, hardware-connected automation, and abusive behavior toward models.
New consolidated section on deceptive activity
Anthropic introduced a new section titled Do Not Engage in Deceptive Campaigns or Artificial Activity that consolidates prior rules spread across elections, fraud, privacy, and disinformation sections. The new text covers deceptive actions of any kind — political or commercial — including obscuring who is behind a message, amplifying content via fake accounts or posts, and creating the tools and infrastructure to run influence operations.
Refocused elections policy
The policy refines political-content restrictions to specifically prohibit using Claude to deceive voters or disrupt elections. The elections section has been renamed Do Not Undermine Democratic Processes and prohibits examples such as spreading false information about candidates or how to vote, impersonating candidates or election officials, and attempting to suppress turnout. Anthropic removed a blanket ban on personalized vote and campaign targeting because it blocked legitimate civic work (for example, nonprofits preparing voter information in other languages or election officials sending ballot cure notices). Behaviors that motivated the original ban — like deceptive targeting or misuse of voters’ personal data — remain prohibited under the deceptive campaigns, surveillance, and privacy sections.
Clearer language on weapons prohibitions
While Anthropic has long prohibited using Claude to develop weapons, the updated policy explicitly states that the ban includes software and components that make weapons operate, as well as actions such as arming drones and other autonomous vehicles. The company says this wording reflects how it has already enforced the policy in practice.
More precise surveillance and law enforcement restrictions
Citing its September threat intelligence report, Anthropic rewrote the surveillance and criminal justice section to clarify prohibited uses. The policy now states that tracking people without their consent is prohibited, whether in real time or through analysis of previously collected data. Claude cannot be used to decide or recommend who to investigate, arrest, or charge in a law enforcement or criminal justice process. The policy also prohibits using Claude to build or improve tools designed for surveillance.
The update notes that certain uses remain permitted when appropriate consent or legal basis exists, including fraud monitoring, content moderation, journalism, and legal research.
High‑risk use case requirements and hardware controls
Anthropic reiterates that applications affecting someone’s health, legal rights, finances, livelihood, or access to essential services require a qualified human in the loop who can review and, if necessary, change Claude’s recommendations, and that affected individuals must be informed that AI was used. While these requirements are unchanged, the High-risk Use Case Requirements section has been rewritten to more clearly list which types of recommendations are covered.
Following the launch of its Model Hardware Standard, the policy also adds requirements for cases where models are connected to hardware that takes autonomous physical actions and might cause injury: a qualified operator must be able to observe the equipment and stop it if needed, and the equipment must be able to hold a safe state if Claude is disconnected.
Prohibition on sustained abusive behavior toward models
Anthropic added a prohibition on sustained and needless abusive or cruel behavior toward its models. The company says this applies only in extreme cases where users repeatedly act cruelly toward the models with no discernible purpose, and does not target ordinary frustration, dark creative themes, or testing and research. Anthropic already allows Claude models to end rare conversations with persistently abusive users on Claude.ai and Claude Code, and ending such interactions remains the primary enforcement mechanism.
Clarifying the Supported Regions Policy
Anthropic updated the Supported Regions page to clarify enforcement: Claude is prohibited for use by people physically located in unsupported regions, by entities incorporated or headquartered in such regions, and by entities majority‑owned or controlled by persons or organizations in those regions.
Looking ahead
Anthropic says it will continue to update the Usage Policy as Claude’s capabilities and risks evolve, seeking input from internal teams, policymakers, subject-matter experts, civil society, and product users. The full Usage Policy and Supported Regions Policy are available on Anthropic’s website; the revised policy becomes effective November 12, 2026.
