About a week ago, the Commerce Department’s Center for AI Standards and Innovation (CAISI) announced an agreement with AI companies Microsoft, xAI, and Google that allowed the government to inspect unreleased AI models before they were made public. Anthropic and OpenAI have signed something similar in 2024.
Here is a long extract from the government’s announcement, dated May 5, 2026:
“Today, the Center for AI Standards and Innovation (CAISI) at the Department of Commerce’s National Institute of Standards and Technology announced new agreements with Google DeepMind, Microsoft, and xAI. Through these expanded industry collaborations, CAISI will conduct pre-deployment assessments and targeted research to better assess AI capabilities at the border and advance the state of AI security. These agreements build on previously announced partnerships, which have been renegotiated to reflect the Secretary of Commerce’s CAISI guidance and the U.S. AI Action Plan.
But this excerpt must have been taken from the Wayback Machine because this announcement is currently missing from the CAISI website. Reuters appears to have been the first to notice, writing on Monday afternoon that using the original URL returned an error page saying “Sorry, we can’t find this page” and then later redirecting to the main CAISI page on the Commerce Department’s website. As of this writing Monday evening, the URL is still a redirect to the CAISI page.
“These agreements support information sharing,” the archived announcement said, while “ensuring a clear understanding within government of AI capabilities and the state of international AI competition.”
Gizmodo requested comment from the White House and the Commerce Department late Monday, but did not immediately receive a response. We will update this article if we receive a response.