Why is NSA using Anthropic's ‘Mythos' despite Pentagon ban? Here's everything to know

The National Security Agency is reportedly using Anthropic’s most powerful artificial intelligence (AI) model, “Mythos Preview,” despite an earlier ban from the Pentagon.

Mythos is labelled as the most powerful AI model of the company yet.

As reported by Axios, sources familiar with the model stated that the development has exposed a sharp divide within the U.S. government.

Presenting an argument in the court proceedings, the Department of Defense argued that Anthropic poses a national security threat.

However, despite this strong argument, the NSA, which operates under the Pentagon’s umbrella, is reportedly among 40 organisations granted access to the cutting-edge model.

But the question remains, why?

NSA made it clear that the primary motive for utilising Mythos is due to defensive cybersecurity, not for spying or offensive hacking.

More specifically, the agency is relying on the model to scan its own networks and environments for exploitable security vulnerabilities. The practice is officially referred to as “hardening” a system.

The model is not a typical AI. Rather, it is powerful enough to flag and exploit a 17-year-old vulnerability in FreeBSD overnight. Additionally, it found “thousands” of zero-day vulnerabilities (flaws unknown to the public) in major operating systems and web browsers.

Not only the NSA, but the rest of the administration also appears eager to engage with Anthropic CEO. 

On Friday, April 17, Anthropic CEO Dario Amodei met with White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent to discuss governmental utilisation of Mythos.