US restores UK early access for AI safety checks
The US has just restored the U.K.'s early access to top AI models from OpenAI, Anthropic, and Google DeepMind, so now British experts can run safety checks before these tools go public.
This change is meant to help both countries work together on making AI safer and spotting problems before they spread.
UK lobbying prompts White House reversal
Last month, the US blocked UK researchers from previewing these AI systems, which made it tough for them to test for risks.
After some lobbying by UK officials (including their AI minister), the White House changed course.
The timing lines up with the U.K.'s push at the G-20 for shared principles and standards for advanced AI and planned changes to Labour's Cyber Security and Resilience Bill, basically aiming to keep innovation exciting but also safe.