Where: United States
When: Reported July 2026
What happened: Reuters sources say CISA is using Anthropic’s Mythos AI model to scan U.S. government code repositories for software flaws that could be exploited by foreign intelligence services or cybercriminals. The reported use reflects growing U.S. government interest in AI-assisted security reviews, even as Anthropic’s relationship with the White House has reportedly remained strained.
CISA reportedly using Anthropic AI to review government software
Edward Kiledjian
@ekiledjian