On May 11 the Commerce Department removed a webpage that had described its agreement with Google, xAI and Microsoft to permit federal testing of new artificial intelligence models for security weaknesses. Visitors following the link that once led to the announcement encountered a message stating "Sorry, we cannot find that page."
After the initial failure to find the page, the same address redirected users to the Center for AI Standards and Innovation's website - the entity identified by officials as the government organization charged with conducting the model vulnerability examinations.
The Commerce Department had announced on May 5 that participating technology companies would provide new AI models to government scientists before those models were deployed to the public. The stated purpose of that early access was to allow federal researchers to test for security flaws that could enable threats such as cyberattacks or military misuse.
U.S. government concern over the national security implications of powerful AI systems has been growing, according to statements accompanying the initiative. Officials referenced that this worry extends to a range of advanced systems, including Anthropic's Mythos, and described early access as a mechanism to surface potential threats.
It remains unclear why the Commerce Department pulled the page from its site. The department and representatives of the Trump White House did not immediately provide comment in response to requests.
The removal came after the announcement that the private firms would hand over models prior to public rollout so government scientists could evaluate them for vulnerabilities. The testing program was framed by officials as a way to identify threats spanning from cyber intrusions to possible military applications.
Market mentions were included in the previously posted material, listing MSFT down 0.77% and GOOGL down 2.76% in the context of broader market indicators that had appeared alongside the announcement. The announcement and subsequent removal highlight ongoing tensions between rapid commercial AI deployment and efforts by government entities to assess systemic security risks before models enter widespread use.
Context and current status
As of the page removal, the public-facing record of the specific agreement and its details are no longer available on the Commerce Department site. The redirect to the Center for AI Standards and Innovation provides a link to the designated government organization tasked with the testing, but the reasons for taking down the original announcement were not provided.
Communications
Requests for explanation were not answered immediately. Spokespeople for the Commerce Department and for the Trump White House did not respond to outreach seeking clarification about the decision to delete the online announcement.