The U.S. government is preparing to make a version of Anthropic PBC’s artificial intelligence model, Mythos, available to major federal agencies as it works around concerns related to cybersecurity risks, according to a memo reviewed by news outlets.
In an email to officials at Cabinet departments, Gregory Barbaccia, the federal chief information officer at the White House Office of Management and Budget, said OMB is developing protections that would permit agencies to begin using the AI tool. Barbaccia’s message framed the move as a preparation of safeguards rather than a confirmation of immediate, broad access.
The communication did not state that agencies will definitively receive access to Mythos, and it offered no schedule for when any deployment might occur. It also did not provide specifics on how agencies would put the model to use or which functions it might support.
According to the memo, top technology and cybersecurity chiefs across Cabinet departments were told to expect additional information in the coming weeks. That follow-up guidance is intended to outline the protections and the conditions under which the AI model could be used, though the memo itself leaves those details unspecified.
The document and the email signal an effort by OMB to balance potential operational adoption of an AI model with attention to cybersecurity concerns. At this stage, the notice serves primarily as advance warning to agency leadership that OMB is working on enabling tools while putting protections in place.
Clear summary
OMB’s federal chief information officer informed Cabinet department officials that the agency is setting up protections that would allow major federal agencies to begin using Anthropic’s Mythos AI model. The email did not confirm agency access, did not give a timeline, and did not describe specific uses. Technology and cybersecurity leaders were told to expect more information in the coming weeks.
Key points
- OMB is preparing protections to enable agency use of Anthropic’s Mythos AI model.
- The email from Gregory Barbaccia does not confirm access, provide a deployment timeline, or enumerate use cases.
- Top technology and cybersecurity chiefs were advised they will receive additional information in the coming weeks.
Risks and uncertainties
- Cybersecurity risks are explicitly cited as a concern prompting the protective measures - this directly affects federal IT and security operations.
- There is uncertainty over whether agencies will obtain access to Mythos, as the email did not confirm distribution or use.
- No timeline or defined use cases were provided, leaving procurement, integration, and operational planning in flux for affected departments and vendors.