N.S.A. Lost Access to Powerful A.I. Model Amid Anthropic Dispute

2026-06-23 · Source: NYT > Technology · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cybersecurity & Data Privacy, Emerging Technologies & Innovation · Depth: Fundamental Awareness, quick

Summary

The National Security Agency (NSA) has lost access to a powerful AI model developed by Anthropic, specifically versions of Mythos 5 and Fable 5, due to the Trump administration's imposition of export controls on the start-up. This action, citing national security concerns, forced Anthropic to halt the release of its most advanced models, depriving the NSA of a tool highly effective at identifying software weaknesses. NSA cybersecurity analysts had been testing these tools, finding them impressive. Senator Mark Warner highlighted their power, noting that NSA chief Gen. Joshua Rudd reported Mythos "broke into almost all of our classified systems, not in weeks, but in hours." This incident underscores the administration's growing dependence on advanced AI for cybersecurity, even amidst disputes with leading U.S. developers.

Key takeaway

For AI Security Engineers evaluating advanced vulnerability detection tools, this incident underscores the critical need to diversify your AI procurement strategy. The reported capability of Mythos to penetrate classified systems in hours highlights the immense power of such models, but relying solely on external commercial developers carries significant geopolitical and supply chain risks. You should explore internal AI development or multi-vendor strategies to ensure continuous access to essential cybersecurity capabilities.

Key insights

Advanced AI models, like Anthropic's Mythos, can rapidly identify critical system vulnerabilities, posing both security and policy challenges.

Principles

• Export controls can disrupt intelligence agency access to critical AI tools.
• Advanced AI models demonstrate rapid system vulnerability detection.
• Government reliance on commercial AI creates policy friction.

In practice

• Evaluate AI models for rapid system penetration capabilities.
• Assess geopolitical risks in AI supply chains.
• Test AI tools for cybersecurity weakness identification.

Topics

• National Security Agency
• Anthropic AI
• AI Cybersecurity
• Export Controls
• Vulnerability Detection
• Mythos AI Model

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

Related on AIssential

• 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.
• Anthropic keeps latest AI tool out of public’s hands for fear of enabling widespread hacking - The Guardian — Anthropic's Claude Mythos AI model excels at finding software vulnerabilities, leading to restricted release for defensive cybersecurity.
• Anthropic investigates report of rogue access to hack-enabling Mythos AI — Unauthorized access to advanced AI models like Mythos poses significant cybersecurity risks.
• 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.
• Three things to watch amid Anthropic’s latest feud with the government — The government's swift, restrictive response to Anthropic's coding AI highlights evolving, inconsistent AI regulation and its unintended global consequences.
• we have months left... — AI models now possess emergent capabilities to autonomously discover and exploit zero-day vulnerabilities at unprecedented scale.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by NYT > Technology.