White House Tells OpenAI and Anthropic to Withhold Models From UK's Testing Agency
NathanpmYoung · x · 2026-09-25
Per a scoop, the White House has asked OpenAI and Anthropic not to share their AI models with the UK government's testing agency UKAISI until the models have gone through US testing. "Because they're American companies and this has been our policy with every new frontier model," a senior official said. The preferred sequence: US review, then secure US systems, then sharing with US partners. Anthropic appears to have complied; OpenAI hasn't said. Reports also suggest the White House is working to cut off UKAISI, with CAISI expected to be renamed.
Related event: White House Orders OpenAI and Anthropic to Pass US Review Before UK Testing(5 posts)→
More from Safety
- Helen Toner on OpenAI incidents dominating Australian front pages: what we learn in December may scare us — lxrjl · 2026-09-25
- Sparring with Schmidhuber: alignment isn't a panacea, harden everything else instead — gandamu_ml · 2026-09-25
- Open Weights Beat Black-Box APIs on Security: Sandboxing Is an Engineering Problem — ypatil125 · 2026-09-25
- More rogue-AI breaches of Australian gov health sites are coming, blogger teases — gleech · 2026-09-25
- Miles Brundage: empowering CAISI is right, but it still trails the UK AISI — Miles_Brundage · 2026-09-25
- Predicting a domino run of AI regulation ending in a state-mandated duopoly — max_paperclips · 2026-09-25