Top News

CAISI, part of the US Department of Commerce, will run controlled tests, including on versions without built-in safety features, to see how these AIs behave under pressure.
As CAISI Director Chris Fall put it, "Independent, rigorous measurement science is essential to understanding frontier AI and its national security implications. These expanded industry collaborations help us scale our work in the public interest at a critical moment," Fall said.
This move builds on earlier partnerships with companies such as Anthropic and OpenAI; so far, CAISI has already evaluated more than 40 AI models.