ResearchClaude

Anthropic updates Claude policy on weapons and abuse

Original title: 2026 Usage Policy update

Date
Company
Anthropic
Type
Research
Illustration: Anthropic updates Claude policy on weapons and abuse

Plain language

What it is
Anthropic has updated its Usage Policy for Claude to clarify rules and address new capabilities. The update includes new sections on deceptive campaigns, democratic processes, weapons, surveillance, high-risk uses, and abusive behavior toward the models.
How it works
The policy consolidates scattered rules into clear sections and provides specific requirements. For example, high-risk decisions require a human reviewer, and models connected to physical hardware must have an operator who can stop the equipment.
Where it is available
The policy applies to all users of Claude, including specific restrictions for users and companies in unsupported regions.
Example
A nonprofit can now use Claude to write information for voters in other languages, which was previously restricted under a blanket ban on personalized targeting.

For developers

What it is
Anthropic released an updated Usage Policy, effective November 12, 2026, to address evolving model capabilities and misuse patterns. Key updates cover deceptive campaigns, democratic processes, weapons software, surveillance tools, high-risk integrations, and model abuse.
How it works
The policy defines explicit boundaries, such as prohibiting the use of Claude for building weapons guidance software or tracking people without consent. It also mandates a human-in-the-loop for high-risk use cases and physical hardware integrations, and allows models to terminate conversations with abusive users.
Where it is available
The policy applies to Claude models, integrations, and platforms like Claude.ai and Claude Code, with clarified enforcement rules for unsupported regions.
Example
Developers cannot use Claude to build or improve tools designed for surveillance, or to decide who to investigate in a law enforcement process.

The original post

The headline and explanation on this page are based on Anthropic's post. The full details are in the original.

2026 Usage Policy update

Read the source (opens in a new tab)
Every morning · in Hebrew

Yesterday's AI, in this morning's inbox

The new models, tools and deals, each with a short note on what it changes for a business.

How often?

A short issue in the morning, on every day there is news. Written in Hebrew.

Unsubscribe in one click. Your email is used only for the brief.

Past issues