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...
The National Security Agency is hacking foreign governments with a corporate AI. It’s adapting Anthropic’s Mythos model for cyber attacks on 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.
Anthropic is still fighting the Pentagon in court. The Defense Department branded the company a supply chain risk for refusing to let its Claude models power mass surveillance or autonomous drone strikes. That principled stand now looks selective.
The firm blocked its AI from one set of government weapons but is now wiring another directly into America’s cyber offense. Pentagon lawyers are battling over drones. Over at Fort Meade, the operation is already live.
Corporate ethics, it turns out, are often just a matter of which government desk you’re sitting at.
Related Reading
Browse all Anthropic news →
Open Source D&B rebuilds 642 million‑business database after AI agents hit limits
Open Source Trump cracks down on Anthropic after Amazon tip; staff largely foreign
Open Source Study: 6.7% of Deepfake Requests on Hugging Face Targeted Children
Related Topic Anthropic says Claude writes >90% of its code; AI pause button urged
LLMs & Generative AI
Related Topic Claude Mythos highlights EU AI safety gaps, says researcher Caroli
Policy & Regulation
From Archives TensorFlow Emotion Dataset with 54,263 Texts Shows Class Imbalance
From Archives xAI launches Grok Imagine 1.5, adding 720p text‑prompted image‑to‑video
From Archives Lovable signs multiyear Google Cloud deal, 5× usage boost, adds Claude, Gemini
Business & Startups
From Archives Trump signs first major AI regulation order of second term, cites Anthropic
Common Questions Answered
How is the NSA using Anthropic's Mythos AI model in cyber operations?
The NSA has adapted Anthropic's Mythos model specifically for conducting offensive cyber attacks against China and Iran. The operation is reportedly already live and represents a direct application of the AI model to government cyber warfare capabilities.
What is the contradiction between Anthropic's stance on Pentagon contracts and its work with the NSA?
Anthropic has been fighting the Pentagon in court, refusing to allow its Claude models to power mass surveillance or autonomous drone strikes on principled grounds. However, the company is now providing its Mythos model directly to the NSA for offensive cyber operations, suggesting its ethical position may be selective depending on the type of military application.
Why did the Defense Department brand Anthropic as a supply chain risk?
The Pentagon classified Anthropic as a supply chain risk because the company refused to let its Claude AI models be used for mass surveillance operations or to power autonomous drone strikes. This refusal to cooperate with certain military applications led to legal battles between Anthropic and Defense Department lawyers.
What is the significance of the NSA's cyber operations being located at Fort Meade?
Fort Meade is the headquarters of the NSA, and the fact that the Mythos AI-powered cyber offensive operations are already live there indicates that the program has moved beyond planning stages into active deployment. This location represents the operational center where the adapted AI model is being used for real offensive cyber capabilities against foreign governments.
Further Reading
- Scoop: NSA using Anthropic's Mythos despite blacklist — Axios
- Anthropic embeds engineers at NSA to deploy Mythos AI model - FT — Investing.com
- Anthropic aids NSA with 'Mythos' to bolster offensive cyber operations — Biz Chosun
- Report Says Anthropic Assists US National Security Agency in Deploying Mythos for Offensive Cyber Warfare — AASTOCKS
