Anthropic Refuses DoW Demands to Remove Claude AI Safety Safeguards

Anthropic

Feb 26, 2026 · Updated Apr 25, 2026

Anthropic is refusing DoW demands to remove Claude safeguards on mass domestic surveillance and fully autonomous weapons. The DoW threatened to invoke the Defense Production Act and label Anthropic a 'supply chain risk' — a first for any American company.

Anthropic CEO Dario Amodei published a statement refusing two DoW demands: removing Claude safeguards for mass domestic surveillance capability and fully autonomous weapons support. The DoW threatened to terminate Anthropic's access, label it a "supply chain risk" - previously reserved for US adversaries - and invoke the Defense Production Act to force the safeguards' removal.

Anthropic argues mass surveillance threatens civil liberties that AI amplifies at scale, and current frontier models aren't reliable enough for weapons that select and engage targets without human oversight. Neither safeguard was ever part of Anthropic's DoW contracts.

Anthropic will enable a smooth transition if offboarded but won't change its position. The DoW's threats are contradictory: one brands Anthropic a security risk, the other treats Claude as mission-critical to national security.

View the full update on anthropic.com

Anthropic

@AnthropicAIFeb 26

A statement from Anthropic CEO, Dario Amodei, on our discussions with the Department of War. https://t.co/rM77LJejuk

9.5k

View on X

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

See all AI news & updates from Anthropic →

Keep reading

Anthropic supports new White House framework for frontier model security review

Anthropic has committed to a new White House Executive Order that establishes a voluntary security review process for advanced AI models. The order prioritizes AI-enabled cyber defense across federal agencies while explicitly prohibiting mandatory government licensing for new model development.

OpenAI Deploys AI in US Classified Networks With Multi-Layer Safety Guardrails

OpenAIFeb 28

OpenAI Deploys AI in US Classified Networks With Multi-Layer Safety Guardrails

OpenAI signed a deal with the US Department of War to deploy AI in classified environments. OpenAI claims stronger guardrails than any prior classified AI agreement — retaining safety stack control, cloud-only deployment, and cleared personnel in the loop.

ClaudeMay 1

Anthropic Launches Claude Security Beta to Automatically Scan and Patch Codebases

Anthropic launched Claude Security in public beta for Enterprise customers to identify and remediate vulnerabilities across entire codebases. Unlike traditional scanners that rely on pattern matching, the tool uses reasoning to trace data flows and validate findings through an adversarial pass. This shift reduces false positive fatigue by ensuring every reported issue includes a verified, human-reviewable patch.