WASHINGTON, D.C. — The Center for AI Standards and Innovation at the Department of Commerce announced on Tuesday that Microsoft, Google's DeepMind and xAI will allow the United States federal government access to their new artificial intelligence models for national security testing. Under the agreement, the U.S. government will evaluate the models before deployment and conduct research to assess their capabilities and security risks.

The agreement fulfills a pledge the administration of U.S. President Donald Trump made in July to partner with technology companies to vet their AI models for national security risks. It builds on 2024 agreements with OpenAI and Anthropic under President Joe Biden's administration, when CAISI was known as the U.S. Artificial Intelligence Safety Institute and focused on developing AI tests, definitions and voluntary safety standards.

"Independent, rigorous measurement science is essential to understanding frontier AI and its national security implications," Center for AI Standards and Innovation Director Chris Fall said. "These expanded industry collaborations help us scale our work in the public interest at a critical moment," he added.

Tom Lue, vice president of global AI affairs at DeepMind, confirmed the partnership in a social media post Tuesday. Elon Musk's xAI, which is not publicly traded, also signed the new agreement to share AI models with the U.S. government.

Microsoft will work with U.S. government scientists to test AI systems in ways that probe unexpected behaviors. The company and government scientists will also develop shared data sets and workflows for testing Microsoft's models. Microsoft signed a similar agreement with the United Kingdom's AI Security Institute.

CAISI has completed more than 40 evaluations, including on cutting-edge models not yet available to the public. Developers frequently hand over versions of their models with safety guardrails stripped back so the center can probe for national security risks. Chris Fall was announced as director of CAISI after ex-Anthropic researcher Collin Burns was pushed out following four days on the job.

In a separate arrangement, the Pentagon and seven tech companies—Google, Microsoft, Amazon Web Services, Nvidia, OpenAI, Reflection and SpaceX—reached an agreement to use their AI systems across classified computer networks. Under that agreement, the companies will provide resources to help augment warfighter decision-making in complex operational environments. Anthropic was not included. Microsoft's stock was down by 0.6 percent in midday trading after the announcement, while Alphabet's stock was up 1.3 percent.