Us Government Expands Vetting Of Frontier Ai Models For Security Risks

Top artificial intelligence firms Microsoft, xAI and Google DeepMind have inked deals with the Commerce Department that will allow the government to vet their AI models for national security risks ahead of release.

The Center for AI Standards and Innovation, an office within the Commerce Department’s National Institute of Standards and Technology, announced Tuesday that it will conduct pre-deployment evaluations and “targeted research” on AI models developed by the companies.

“Independent, rigorous measurement science is essential to understanding frontier AI and its national security implications,” CAISI Director Chris Fall said in a statement.

The agreements closely resemble voluntary vetting deals made in August 2024 between the Biden administration and leading AI labs OpenAI and Anthropic — arrangements that CAISI said have since been "renegotiated" to reflect Trump administration priorities.

They come as the U.S. government moves with increasing urgency to address security risks posed by frontier AI tools, including cybersecurity concerns raised by Anthropic's new Mythos model. The New York Times reported Monday that the White House is considering an executive order to establish a formal government review process for new AI models.

In a statement, CAISI said its deals with Microsoft, xAI and Google DeepMind will “support information-sharing, driving voluntary product improvements and ensuring a clear understanding in government of AI capabilities and the state of international AI competition.”

Spokespeople for Microsoft and Google DeepMind pointed to social media posts from company executives confirming their participation in CAISI’s voluntary vetting regime. According to Natasha Crampton, head of Microsoft’s Office of Responsible AI, the company inked a similar deal with the United Kingdom's AI Security Institute.

Spokespeople for xAI did not respond to a request for comment.