Commerce Department removes AI security testing details from website
The U.S. Commerce Department deleted information about its agreement with Google, Microsoft, and xAI to test artificial intelligence models for security vulnerabilities, according to a review of the agency's website.
The announcement page, originally posted May 5, no longer exists. Visitors to the link now see an error message: "Sorry, we cannot find that page." The link later redirected to the Center for AI Standards and Innovation, the government organization running the tests.
Under the original agreement, the three companies would provide new AI models before public release, allowing government scientists to identify security flaws. Officials said the early access aimed to detect threats ranging from cyberattacks to military misuse.
The Commerce Department and White House did not respond to requests for comment about why the page was removed.
What the testing program covered
The May 5 announcement outlined a voluntary program where tech companies submit advanced AI systems to federal review. The government cited growing concern over national security risks from powerful AI systems, including Anthropic's Mythos.
For government employees working on AI policy or security, understanding how these testing frameworks operate is essential. Learn more about AI for Government initiatives and how they affect your work.
Those developing policy around AI systems may find value in the AI Learning Path for Policy Makers, which covers security vulnerabilities and government oversight mechanisms.
Open questions
The deletion raises questions about the status of the testing program and whether the agreement remains active. Neither the Commerce Department nor the White House provided clarification as of Monday afternoon Washington time.
Your membership also unlocks:
