The NSA is using Anthropic's most powerful AI model Mythos

2026-04-20 · Source: The Decoder · Field: Government & Public Sector — Public Safety & Security, Public Policy & Governance, Artificial Intelligence & Machine Learning · Depth: Fundamental Awareness, quick

Summary

The U.S. National Security Agency (NSA) is reportedly utilizing Anthropic's "Mythos Preview," its most advanced AI model, despite the Pentagon classifying Anthropic as a security risk and attempting to block it as a vendor since February. The Pentagon, which oversees the NSA, has even asserted in court documents that Anthropic's tools threaten national security. Anthropic has limited Mythos access to approximately 40 organizations under "Project Glasswing," citing the model's dangerous offensive cyber capabilities. Anthropic CEO Dario Amodei recently met with White House officials to discuss broader government deployment, but the company has resisted Pentagon demands to make Claude available for "all legal purposes," specifically refusing mass surveillance and autonomous weapons applications. The UK's intelligence services also have access to Mythos via the country's AI Security Institute.

Key takeaway

For government agencies and defense contractors evaluating advanced AI models, be aware that internal security classifications and ethical use policies can create significant friction. Your teams should carefully assess vendor security postures and model capabilities against national security guidelines, while also considering the vendor's willingness to restrict use cases like mass surveillance or autonomous weapons, which could impact deployment timelines and operational scope.

Key insights

The NSA is using Anthropic's powerful AI model "Mythos" despite Pentagon security concerns.

Principles

• AI model access can be highly restricted.
• National security agencies prioritize advanced AI.
• AI developers can set ethical use boundaries.

In practice

• Evaluate AI models for offensive cyber capabilities.
• Implement strict access controls for powerful AI.
• Engage with government on AI deployment policies.

Topics

• Anthropic Mythos
• National Security Agency
• Pentagon Security Concerns
• Offensive Cyber Capabilities
• AI Model Restrictions

Best for: CTO, VP of Engineering/Data, Director of AI/ML, Policy Maker, AI Security Engineer, AI Ethicist

Related on AIssential

• The White House rethinks its Anthropic fight — National security interests are reshaping government policy on advanced AI model access and deployment.
• Cybersecurity experts blast US government for restricting Anthropic’s AI models — The US government's ban on Anthropic's AI models is criticized by experts for hindering defense capabilities and US AI competitiveness without sufficient justification.
• NSA spies are reportedly using Anthropic’s Mythos, despite Pentagon feud — Anthropic's restricted Mythos model is being used by the NSA despite prior DoD "supply-chain risk" designation.
• Mythos... I can't sleep — Anthropic's "Mythos" model poses a cybersecurity threat by autonomously discovering thousands of high-severity software vulnerabilities.
• BREAKING: Anthropic’s “Dangerous” AI Model Mythos LEAKED — Controlling access to powerful AI models is as critical as the models' inherent safety.
• Anthropic has crossed a line — Anthropic's Mythos model offers a step-change in AI capability, posing both significant cybersecurity risks and powerful defensive tools.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.