US Military Reverses Sept. 1 Anthropic Ban for Contractors as China Race Drives AI Push
Updated
Updated · The New York Times · Aug 16
US Military Reverses Sept. 1 Anthropic Ban for Contractors as China Race Drives AI Push
3 articles · Updated · The New York Times · Aug 16
Summary
Within a month of a mid-July warning, military contractors were told they could ignore an Air Force order to remove Anthropic software from weapons and control systems by Sept. 1.
Anthropic’s Mythos model was deemed too important to drop: senior NSA officials argued it helps find and patch Pentagon vulnerabilities, and officials said tests were underway on its ability to penetrate hardened foreign networks.
The reversal exposed a split inside the national security establishment, where Defense Secretary Pete Hegseth has publicly feuded with Anthropic CEO Dario Amodei over limits the company sought on military use.
For Pentagon officials racing to keep pace with China in AI, cutting off Mythos was seen by one recently departed senior U.S. official as risking 'unilateral disarmament' in a fast-moving cyber arms race.
How will defense contractors survive the legal chaos of conflicting Pentagon mandates regarding the world's most powerful cyber AI?
With AI finding critical bugs faster than humans can patch them, is the Pentagon's reliance on commercial tech creating a massive new vulnerability?
If an AI can hack classified systems in hours, why did the military secretly reverse its ban on the very same technology?
Inside the 2026 Pentagon-Anthropic AI Crisis: Policy Whiplash, Legal Battles, and the Future of Military AI Governance
Overview
In 2026, the Pentagon ordered a rapid purge of Anthropic’s AI from military systems, only to reverse course weeks later after the NSA warned that losing access to Anthropic’s powerful Mythos model would cripple defense capabilities. This exposed deep divisions: Anthropic’s CEO refused to allow its AI for mass surveillance or autonomous weapons, while the Pentagon demanded full control. The standoff escalated to federal blacklisting, lawsuits, and industry protests, shaking trust in government tech procurement. Ultimately, Congress responded with new legislation to set clear safety and human-control rules for military AI, aiming to balance national security with ethical boundaries.