The National Security Agency (NSA) is reportedly utilizing Anthropic’s Mythos Preview, a powerful AI model capable of identifying software vulnerabilities, despite recent tensions between the company and the Department of War over supply chain risks. Two months ago, the Pentagon threatened to blacklist Anthropic due to its refusal to remove ethical guardrails for military applications, which led to a lawsuit from the company. However, following a White House directive encouraging federal agencies to adopt Mythos for defensive cybersecurity, the NSA appears to be integrating the model into its operations, while the UK’s AI Security Institute has also secured access to it.

Anthropic: Anthropic is an AI research company focused on developing safe, reliable, and interpretable AI systems, including the Claude family of models. It recently introduced Mythos Preview, a frontier model capable of autonomously discovering zero-day vulnerabilities, which it has restricted to select organizations due to its advanced offensive cyber potential. In this news, Anthropic’s technology is being adopted by the NSA amid evolving government relations following prior disputes with the Department of War.
Susie Wiles: Susie Wiles is the White House Chief of Staff and the first woman to hold the position, with extensive experience as a political consultant. She engaged in recent discussions with Anthropic CEO Dario Amodei regarding the integration of the Mythos AI model into government operations.
Dario Amodei: Dario Amodei is the CEO and co-founder of Anthropic, leading efforts to advance AI capabilities with an emphasis on safety and ethical considerations. He recently met with White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent to discuss deploying the Mythos model within the government and reviewing security practices.
National Security Agency: The National Security Agency is the United States intelligence agency responsible for signals intelligence, cybersecurity, and protecting national communications networks. It is reportedly using Anthropic’s Mythos Preview, the company’s most powerful AI model to date, to scan its environments for exploitable security vulnerabilities. This usage persists despite the Department of War designating Anthropic as a supply chain risk.

`json
{
“Model Capabilities”: “Anthropic’s Mythos Preview is used for discovering security vulnerabilities within organizations.”,
“Government Directive”: “The White House has directed federal agencies to adopt Anthropic’s Mythos for cybersecurity-related purposes.”,
“International Collaboration”: “The UK’s AI Security Institute has access to Mythos.”
}
`