White House Keeps Its AI Rulebook Secret as Companies Seek Answers

White House Keeps Its AI Rulebook Secret as Companies Seek Answers
The White House has built a new gatekeeping system for the most powerful AI models — then kept the rulebook largely out of public view. For companies invited inside, it offers a path to government review; for everyone else, it leaves crucial questions unanswered.

The process began with President Donald Trump’s June executive order, which directed officials to create a benchmark for advanced AI cyber capabilities. The administration’s stated approach remains voluntary, but the envisioned review could give federal officials access to unreleased models for as long as 30 days, under rules governing confidentiality, cybersecurity, insider risk, intellectual property and nondisclosure.

By Monday, the White House said it had met the deadline. “The voluntary framework outlined in the June 2nd executive order was complete by the deadline,” an official said, while confirming that discussions with industry were continuing. The official defended the decision not to publish the document: “Just because things are unclassified that doesn’t mean we are going to broadcast them to everyone.”

On Tuesday, staff-level meetings brought Anthropic, OpenAI, Google and other industry participants into the process. The framework reportedly focuses on closed-source frontier models with leading-edge capabilities and potential national-security risks, while excluding open models. Yet neither “state-of-the-art” nor “national security risk” has a clear public definition.

That split captures the administration’s balancing act. It wants earlier visibility into models that could sharpen hacking or cyber capabilities without imposing a conventional regulatory regime. But privacy has become a policy choice of its own: uninvited companies, researchers, lawmakers and U.S. allies are left to guess which systems qualify, who counts as a trusted partner, and what happens if a review detects a dangerous capability.

The White House says it is engaging “many more” partners than the three best-known labs. Still, the framework’s first practical effect may be to create two AI industries: those briefed on the rules, and those waiting outside the room.

Continue reading https://foxvector.com/stories/019fd04e-fe6c-1eee-72bb-06a1bca6a633