Pentagon CTO Confirms Anthropic Blacklist, Distinguishes Mythos AI as Separate Security Concern

The Pentagon's Chief Technology Officer, Emil Michael, reaffirmed on Friday that Anthropic remains blacklisted as a supply chain risk, but emphasized that the company's AI model Mythos represents a distinct national security challenge. This clarification comes amid ongoing tensions between the Department of Defense (DOD) and Anthropic over the use of AI technologies.

Michael highlighted that Mythos, an advanced AI with cyber vulnerability detection and patching capabilities, requires government-wide attention to secure networks. Meanwhile, the DOD has forged partnerships with seven other AI firms to deploy their technologies on classified networks, signaling a strategic approach to AI integration in defense.

Anthropic Remains a Supply Chain Risk Amid Pentagon Dispute

Earlier this year, the Department of Defense publicly designated Anthropic as a supply chain risk, citing concerns that its AI technology could threaten U.S. national security. The designation followed failed negotiations over how Anthropic's AI models could be utilized by the military. As a result, defense contractors are required to certify that they do not use Anthropic's Claude models in their work with the military.

In response, Anthropic filed a lawsuit against the Trump administration in March seeking to overturn the Pentagon's blacklist. Despite this, Emil Michael confirmed that the blacklist remains in place, with the DOD insisting on enforceable guardrails for AI use, which are subject to negotiation with various companies.

Mythos AI: A Separate National Security Challenge

Michael described Mythos, Anthropic's AI model equipped with advanced cyber capabilities, as a 'separate national security moment.' Unlike the broader supply chain concerns with Anthropic, Mythos specifically focuses on identifying cyber vulnerabilities and facilitating their patching, necessitating heightened network security measures across government agencies.

"The Mythos issue that's being dealt with government-wide, not just at the Department of Defense, is a separate national security moment where we have to make sure that our networks are hardened up," said Emil Michael.—Emil Michael, Department of Defense CTO

DOD Partners with Leading AI Firms for Classified Network Deployment

On the same day, the DOD announced agreements with seven prominent AI companies to deploy their technologies across the agency's classified networks for lawful operational use. These companies include Google, OpenAI, Nvidia, Microsoft, Amazon Web Services, SpaceX (merged with Elon Musk's xAI), and Reflection, a startup developing open weight models.

• Google
• OpenAI
• Nvidia
• Microsoft
• Amazon Web Services
• SpaceX (merged with xAI)
• Reflection (startup)

This move reflects the Pentagon's strategic effort to integrate cutting-edge AI technologies while maintaining strict security protocols, contrasting with its cautious stance on Anthropic's offerings.

Looking Ahead: Balancing Innovation and Security in AI

The ongoing debate over Anthropic's blacklist and the separate handling of Mythos underscores the complex challenges the U.S. government faces in adopting AI technologies. Ensuring national security while leveraging AI's potential requires nuanced policies and collaboration with industry leaders.

As AI continues to evolve rapidly, the Pentagon's approach may serve as a blueprint for balancing innovation with risk management, particularly in sensitive defense applications.