CISA Locked Out of Anthropic's Mythos Preview Model
America's top cybersecurity agency can't access Anthropic's latest AI model, even as other government agencies already use it.
The Cybersecurity and Infrastructure Security Agency — the federal body literally responsible for defending U.S. digital infrastructure — doesn't have access to Anthropic's new Mythos Preview model. That's a problem, given other government agencies are already using it.
Anthropic says it briefed CISA ahead of the Mythos unveiling, but a briefing and actual access are very different things. The nation's cyber defense agency is effectively flying blind on a powerful AI model already deployed across parts of the federal government.
The gap raises obvious questions about how cutting-edge AI tools get distributed within government. If CISA's job is to assess and mitigate digital threats, leaving them without hands-on access to a model other agencies rely on seems like a notable oversight.
Anthropic hasn't detailed why CISA remains on the outside looking in.