The NSA is using Anthropic's Mythos model despite prior conflicts with the Pentagon over surveillance and weaponization issues.
The National Security Agency (NSA) is reportedly using Anthropic's Mythos model, despite the startup's ongoing disputes with the Pentagon, according to sources cited by Axios.
These tensions stem from Anthropic's refusal to permit its AI technologies for mass surveillance of U.S. citizens and the creation of autonomous weaponry. This led to an order from former President Donald Trump requiring federal agencies to stop using Anthropic's technologies within six months.
In April, Anthropic introduced a new AI model, Claude Mythos, but chose not to make it publicly available due to serious security concerns. Testing revealed that the model could identify thousands of zero-day vulnerabilities in major software systems.
Rather than a public launch, Anthropic initiated Project Glasswing, collaborating with major companies like AWS, Google, and Microsoft to test the technology in secure settings.
Axios has learned that the NSA is one of the undisclosed recipients of this technology. While the specifics of the NSA's application remain unclear, other entities are reportedly using it to assess their systems for vulnerabilities.
“Currently, the military is expanding its use of Anthropic's tools while simultaneously claiming in court that their use poses a threat to U.S. national security,” journalists have noted.
Despite earlier conflicts, Anthropic's relationship with government agencies appears to be improving. CEO Dario Amodei met on April 18 with key administration officials, including White House Chief of Staff Suzy Wiles and Treasury Secretary Scott Bessent, to discuss potential collaborations.
Additionally, in April, Anthropic launched Claude Managed Agents, a platform designed for executing complex agent tasks.
