The White House now says its AI review framework was done on time — it's just secret
A June executive order gave the US government 60 days — until August 1 — to stand up a voluntary pre-release review process for frontier AI models, with OpenAI, Anthropic, Google, Microsoft and xAI agreeing to take part. The deadline passed with no framework made public, even as the three biggest labs jointly submitted revisions to the draft and Sam Altman visited the White House. (Correction: on August 3 the White House said the framework was in fact completed by the deadline but will not be published — its testing benchmarks are classified — and invited staff from the big AI labs to review it behind closed doors.)
Why it mattersThis would be Washington's first concrete mechanism for reviewing powerful AI models before they ship. The secrecy — and the fact that the biggest labs are helping write the rules they will be judged by — shows how hard it is to regulate an industry that moves faster than its regulators.
Read in the app — free, in 9 languages
Related stories
Nevada just cleared up to 8,000 driverless taxis for Las Vegas — Tesla's ceiling went from 10 to 5,000 in five weeks2026-08-21OpenAI has started a team whose stated worry is that a company like OpenAI ends up with too much power
2026-08-21OpenAI says it can police misuse without keeping your company's data — and it is pointing at Anthropic while saying so
2026-08-20A Florida officer ran his ex-wife's plate 717 times. A Kansas city has just switched its cameras off.
2026-08-20Six months ago Hollywood sent ByteDance a cease-and-desist. This week they signed a truce — and no money changed hands.
2026-08-20