Microsoft, Alphabet-owned Google, and Elon Musk’s xAI will give the U.S. government early access to new artificial intelligence models before their public release, allowing officials to check for national security risks.
The agreement, announced Tuesday by the Center for AI Standards and Innovation at the Department of Commerce, lets the agency evaluate the models before deployment and conduct research to assess both capabilities and security risks.
CAISI Director Chris Fall said in a statement: "Independent, rigorous measurement science is essential to understanding frontier AI and its national security implications."
The deal builds on previous agreements with OpenAI and Anthropic, established in 2024 under the Biden administration. CAISI has already completed more than 40 evaluations, including on cutting-edge models not yet available to the public. Developers often hand over versions with safety guardrails stripped so the center can probe for vulnerabilities.
Last week, the Pentagon announced agreements with seven AI companies to deploy their advanced capabilities on the Defense Department’s classified networks, broadening the range of AI providers working across the military. That announcement did not include Anthropic, which has been in a dispute with the Pentagon over guardrails on military use of its tools.