The White House reverses a prior policy that required U.S. approval before AI companies share newly released models with British safety testers. Multiple outlets report that this change restores the U.K.’s access to pre-release versions of models from leading AI developers, enabling evaluation ahead of public release.
The reports frame the reversal as a shift in the Trump administration’s approach to international AI safety cooperation. Several sources say the earlier requirement limited coordination with the U.K., while the new stance gives U.S. developers more latitude to work with British policymakers and testers. Coverage also links the change to broader debates over whether AI safety rules should be binding, with some outlets noting that the U.S. position has been cautious about mandatory requirements.
While details of the implementation are not consistent across reports, the core point is that the U.S. now allows model-sharing for pre-release assessment in the U.K., reversing a previous approval step and signaling a more cooperative posture on cross-border AI oversight.
