Despite being labeled a “supply chain risk” by the Pentagon earlier this year, Anthropic is reportedly seeing its newest and most secretive model, Mythos Preview, being used by the National Security Agency (NSA). So, it seems that not even a legal battle and a presidential blacklist can stop the U.S. government from wanting the latest AI tech. A recent report already revealed meetings between authorities and the AI startup related to the new AI model.
Why the NSA is secretly using Anthropic’s Mythos AI model
The situation is a bit of a head-scratcher. Back in February, the Trump administration ordered government agencies to stop using Anthropic’s services. This move followed a breakdown in contract talks where Anthropic refused to lift safeguards that would prevent its AI from being used for autonomous weapons or mass domestic surveillance. There is an active deadline of six months for compliance with the order.
However, according to a report by Axios, the NSA is currently one of about 40 organizations granted exclusive access to Mythos. Sources suggest the model is being used even more widely across various departments. Tjis reflects a clear divide between the government’s public stance and its operational needs.
What is Mythos Preview?
Anthropic has been quiet about the specifics of Mythos. However, the company describes it as a general-purpose model with a “striking” talent for computer security. The company even restricted its rollout to a very small, vetted group due to its powerful offensive cyber capabilities. Cybersecurity initiative Project Glasswing provides early access to the power of Mythos to several key names.
The NSA’s exact activities remain classified. That said, other organizations with access to Mythos are primarily using it to scan their own digital environments for vulnerabilities. Essentially, the model acts as a high-tech security guard that can find the cracks in a system before a hacker does.
Meetings at the White House
The timing of this leak is particularly interesting. Just last Friday, Anthropic CEO Dario Amodei met with White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent. While President Trump told reporters he had “no idea” the meeting took place, officials described the talks as “productive and constructive.”
The discussion reportedly centered on how government agencies—outside of the feuding Pentagon—can safely engage with Anthropic’s tools. The administration is still looking for a way to utilize the company’s tech without compromising on their respective red lines.
The U.S. isn’t the only one getting a head start. Across the pond, the U.K. AI Security Institute has also confirmed it has access to Mythos.
The post NSA Reportedly Using Anthropic’s Mythos AI Despite Government Blacklist appeared first on Android Headlines.
