Washington's secret AI review framework covers only closed models
New reporting on the Trump administration's completed but unpublished frontier-AI review framework says it defines a covered model as closed-source, state-of-the-art and posing national security risks — open-weight models are explicitly excluded, and nothing in the framework restricts them once released. The White House walked industry representatives through the details on August 4 but still won't make the rules public, and key terms like 'state of the art' remain undefined.
Why it mattersWhere this line sits decides who gets reviewed before release. With Chinese and American open-weight releases booming, the fastest-growing part of the ecosystem stays outside the government's new gate — a deliberate bet that open models are an asset, not a threat.
▶ Related video: Understanding The Trump Executive Order on Frontier Models in AI
Read in the app — free, in 9 languages
Related stories
Norway moves to ban AI glasses from playgrounds and changing rooms.2026-10-06Under oath, four AI firms would not say if AI could be catastrophic.
2026-10-06Anthropic handed a user's diary entry to the police.
2026-10-05Two senators want prison time for bosses whose AI agents hack.
2026-10-05Musk's xAI got America's first AI nudify ban frozen in court.
2026-10-05