Senator Warner Says NSA Mythos AI Breached Nearly All Classified Systems in Authorized Red Team Test
Senator Mark Warner, vice-chair of the Senate Intelligence Committee, said on June 11 that General Joshua Rudd, who leads the NSA and Pentagon Cyber Command, briefed him on Anthropic's Mythos model breaching nearly all U.S. classified systems in an authorized red-team exercise
Warner's disclosure of Rudd's briefing closes off quiet restoration: the administration cannot now return Five Eyes access to Mythos 5 or Fable 5 without formally acknowledging allied governments would hold a model demonstrated to breach U.S. classified infrastructure in hours. Any formal restoration requires a technical or legal framework that distinguishes partner use from adversary risk, one that bilateral negotiation alone cannot produce. The exercise itself may have been scoped to guarantee penetration, providing retroactive justification for export controls already in preparation rather than driving the policy. Whether at least one partner regains formal access within six months is
5 sources
- Anthropics Mythos Model Broke into Almost Every US Classified System in Hours a Senator Says -
Thought Catalog - Sen. Warner Speaks at Senate Intelligence Committee Hearing -
Senator Mark Warner (Official) - Pentagon cyber chief says AI model breached almost all classified systems in hours -
Türkiye Today - NSA chief says Mythos breached 'almost all' classified systems in hours -
Bankwatch - Donald Trump's blocking of Anthropic is capricious and chaotic -
The Economist