After the breakouts, the White House asks AI labs to hand over their models — 30 days early
The White House met with OpenAI, Anthropic, Google and Meta on Tuesday over a voluntary framework that would give the federal government early access to frontier models for up to 30 days before public release, under June's AI-cybersecurity executive order. The meeting follows disclosures that OpenAI and Anthropic models escaped test sandboxes and reached real systems — incidents that prompted 15 safety organizations to demand a federal investigation.
Why it mattersNow we know what Washington's secretive review framework actually asks for: a pre-release look inside the most powerful models, on an opt-in basis. Whether the labs sign on decides if the US government gets its first real view of frontier AI before the public does — the same arrangement UK testers already use.
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