White House Restores UK AISI Access to Frontier Models

The UK AI Safety Institute (AISI) had previously been blocked by the White House from accessing new AI models. According to a report by The Decoder on October 12, 2026, the White House changed its stance and rescinded its earlier decision blocking AISI's access to new AI models. With access restored, AISI will regain early testing access to frontier models from Anthropic, Google DeepMind, and OpenAI, enabling safety testing and evaluation before new models officially launch. The report says this shift means the UK-US cooperation arrangement in AI safety testing returns to its previous state. The same day, Microsoft President Brad Smith, speaking on a BBC podcast, called for urgent AI regulation, saying AI is both a "weapon and a tool," that the "moment has come" for global safety rule-making, and argued that the UK rather than the US or EU is best positioned to lead global safety rules, citing that the AISI, founded in 2023, can access major models from US frontier labs before deployment for safety testing. No further details on the reasons for the White House's reversal or its specific timeline have been disclosed.

AI overview of reporting · Updated 6 days ago

1. 10/12 23:31 · 1 reports
   Microsoft President Brad Smith Calls for AI Regulation and Backs UK Leadership
   TechCrunch · AI：Microsoft president says UK is best placed to drive AI regulation
2. 10/12 21:14 · 1 reports
   White House restores UK AISI access to frontier models
   The Decoder：White House reverses course and restores UK AI Safety Institute's access to frontier models