Report: White House asked OpenAI and Anthropic to withhold models from UK AISI until US review is complete
Following·Day 2·first covered 24 Sep 2026·2 sources
White House Asks Labs to Withhold Models from UK AISI Pending US Review
The gist
Pre-deployment access to frontier models is central to how safety institutes evaluate risks before public release. Restricting that access to the UK AISI puts its evaluation work in question at a time when its US counterpart is, by one researcher's account, less capable.
The full picture
The White House asked OpenAI and Anthropic not to share new models with the UK's AI Security Institute until a US review is complete, raising concerns about the AISI's ability to conduct pre-deployment evaluations. Separately, AI safety researcher Miles Brundage said the US CAISI currently lags behind the UK AISI and warned that moves to empower it could ironically leave the US government less informed about AI risks. UK Prime Minister Andy Burnham called for a global effort to contain AI threats and described Britain as an 'honest broker' between the US and China during its G20 presidency.
How it developed
Miles Brundage posted concern that US CAISI trails UK AISI and restructuring could leave US gov less informed
Sources
Miles Brundage Twitter
Transformer News
Related
The daily email
Want this in your inbox?
I send one email each morning with the stories that moved. If you would rather just read here, that works too.
Subscribe free