WASHINGTON, May 12 — The U.S. Commerce Department has removed from its website details about an agreement with major artificial intelligence companies, including Google, xAI and Microsoft, to test advanced AI models for potential security vulnerabilities.
According to a Reuters review, the webpage that previously hosted the announcement now displays an error message stating, “Sorry, we cannot find that page.” Later, the deleted link redirected users to the website of the Center for AI Standards and Innovation, the government body overseeing the evaluations.
The Commerce Department had announced on May 5 that participating AI companies would provide government scientists with early access to upcoming AI models before public release. The goal was to allow officials to identify possible security flaws and assess risks tied to increasingly powerful AI systems.
The move came amid rising concern within the U.S. government about the national security implications of advanced artificial intelligence technologies. Officials have warned that sophisticated AI models could potentially be exploited for cyberattacks, disinformation campaigns, or even military misuse.
Concerns have also intensified around emerging systems such as Mythos from Anthropic, which has drawn attention for its advanced reasoning capabilities.
It remains unclear why the Commerce Department removed the webpage or whether the testing agreement itself has changed. Neither the department nor the companies involved immediately commented on the matter.