AI Companies Get to See Their Own Safety Scores. The Public Doesn't.

The White House finalized an AI safety testing framework — then decided only the companies being tested would see the results. That's not oversight. It's a liability shield built with industry input.