OpenAI · Anthropic · Safety and security
The US is again letting AI companies share their models with Britain’s AI Security Institute before release
The Trump administration has dropped its request that US developers such as OpenAI and Anthropic hold new models back from Britain’s AI Security Institute (AISI) before release, The Telegraph reported, as relayed by Reuters on October 11. The request, first reported by Politico on September 24, came from the White House Office of the National Cyber Director, which wanted US officials to review the models first. Anthropic did not share Claude Mythos 5.1 with the institute in the meantime. On October 12, UK AI minister Kanishka Narayan said Britain had worked with Washington and the labs to keep AISI’s pre-release access, but warned that no outcome is permanent when relying on another country’s technology. The White House, OpenAI and Anthropic did not comment.
Sources
- Reuters (Global Banking & Finance Review üzerinden), “US reverses course on Britain’s testing of AI models, Telegraph says”, Reuters wire relaying The Telegraph, (globalbankingandfinance.com)
- GOV.UK, “Building a Safer, More Secure Britain in the age of AI”, Narayan: AISI keeps pre-release access, (gov.uk)
- The Next Web, “White House asks OpenAI and Anthropic to hold AI models from UK testers”, Background: the National Cyber Director's request, (thenextweb.com)
- The Decoder, “White House reverses course and restores UK AI Safety Institute’s access to frontier models”, Access restored, (the-decoder.com)
- Quartz (Yahoo News üzerinden), “U.K. regains access to OpenAI and Anthropic AI models pre-release”, Narayan's caution on permanence, (yahoo.com)
About this story
This story was posted on Instagram by @jarrus.tech on Oct. 13, 2026.
Spotted an error in this story? [email protected] · Instagram
Short link: thejarrus.com/en/us-uk-model-testing
This story in Turkish: ABD, yapay zeka şirketlerinin modellerini yayından önce İngiltere’nin Yapay Zeka Güvenlik Enstitüsü’yle paylaşmasına yeniden izin verdi
On the same topic
CNBC: Small groups like METR, Apollo Research and Transluce are emerging as AI’s safety gatekeepers
AI labs are leaning on independent evaluators, CNBC reported, but there are still no standards for funding, access or reporting.
Only 3.6% of Chinese AI model releases came with a published safety test result
Only 31 of 857 model releases from nine major Chinese developers came with a published safety evaluation result, SemiAnalysis found.
AI pentesting tool used in Korean bank attacks pulled from public release
The developer of ARTEX pulled it from public release, citing misuse; experts say copies already downloaded remain usable.