Anthropic’s Mythos model found vulnerabilities in classified US government systems, official says
Anthropic’s Mythos model found vulnerabilities in classified US government systems, official says
Publish Date: 2026-06-23 20:20:00
Source Domain: www.castanet.net
- Anthropic’s AI model, Mythos, identified vulnerabilities in highly secure U.S. government systems during a testing exercise conducted as part of Project Glasswing.
- Although Mythos discovered these vulnerabilities in hours, it did not mean there was an immediate threat, as it did not exploit them.
- Tensions exist between Anthropic and the Trump administration over the use and accessibility of AI models due to security concerns.
- The Trump administration issued a directive to restrict the use of Anthropic’s advanced AI models by foreign nationals, arguing that national security risks may arise in their use.
- A group of cybersecurity experts argued against the directive, claiming it hampers U.S. cyber defense capabilities and potentially aids adversaries.
