NSA leveraging Anthropic’s Mythos AI model for offensive cyber operations targeting Chinese and Iranian cyber threats, illust

Editorial illustration for NSA uses Anthropic's Mythos AI model for offensive cyber ops against China, Iran

NSA uses Anthropic's Mythos AI model for offensive cyber...

NSA uses Anthropic's Mythos AI model for offensive cyber ops against China, Iran

By AI Daily Post Edited by Brian Petersen, Editor-in-Chief

June 5, 2026 • Updated: June 7, 2026 • 2 min read

The Financial Times says the NSA is now running offensive cyber operations with Anthropic’s Mythos AI model. While Anthropic has placed roughly half a dozen engineers on‑site to adapt the system, it’s not clear whether those staff are directly involved in any strikes against networks in China or Iran. The move arrives as the company battles a Pentagon‑led legal fight; the Department of Defense labelled Anthropic a “supply chain risk” and tried to bar it from contracts after the firm pushed back on using Claude models for mass surveillance and autonomous drones.

The NSA delivery, however, was exempt from that ban. Anthropic argues its stance aligns with its own values, repeatedly citing the need to protect U.S. citizens from AI threats.

Meanwhile, the startup has broadened Mythos access to 150 organizations in 15 countries, and it welcomed President Trump’s executive order on voluntary safety testing of new AI models. The juxtaposition of a commercial AI firm and a national intelligence agency raises questions about how far the line between private tech and offensive cyber tools can be drawn.

Anthropic's Mythos model is reportedly powering NSA offensive cyber ops against China and Iran The US National Security Agency is using Anthropic's Mythos AI model for offensive cyber operations, the Financial Times reports. Anthropic has placed about half a dozen engineers directly at the NSA to adapt the model and support its use. Mythos could be used to break into networks in China or Iran.

Whether Anthropic's engineers are involved in active operations isn't clear. The move comes while Anthropic is still in a legal fight with the Pentagon. The Department of Defense classified the company as a "supply chain risk" and tried to cut it from contracts because Anthropic wanted to restrict the use of its Claude models for mass surveillance and autonomous drones.

Anthropic's Mythos model is reportedly powering NSA offensive cyber ops against China and Iran - THE DECODER

Why this matters

How do we interpret a leading AI firm embedding engineers inside a national intelligence agency? Anthropic has stationed roughly six engineers at the NSA to adapt its Mythos model for offensive cyber operations targeting China and Iran, according to the Financial Times. The model’s alleged capability to breach foreign networks raises immediate questions about dual‑use risk and the adequacy of internal controls.

While the report confirms the partnership, it does not clarify whether Anthropic’s staff are directly participating in active intrusions, leaving a gap in accountability. A sobering reminder. For developers, the episode underscores that powerful generative models can be repurposed beyond commercial aims, prompting us to revisit security‑by‑design practices.

Founders may need to anticipate government contracts that blur ethical lines, and researchers should expect heightened scrutiny of their work’s potential weaponization. Yet, the precise boundaries of Anthropic’s involvement remain uncertain, and we lack visibility into any oversight mechanisms the agency may have instituted. As we've continued to build increasingly capable systems, the line between innovation and misuse demands careful, ongoing attention.