Telegraph reports U.S. reversal on early AI access for British testers

The Trump administration has withdrawn its request that U.S. developers withhold unreleased AI models from Britain’s AI Security Institute, The Telegraph reported on October 11.

According to the newspaper’s account, companies including OpenAI and Anthropic can again provide the British government with early access for safety testing without conflicting with the administration’s position.

The institute was established after the international AI safety summit in 2023. Its work includes evaluating advanced models’ capabilities and the risks that those systems could be misused.

Spokespersons for Trump, OpenAI and Anthropic did not immediately respond to Reuters’ requests for comment. The reported change therefore remains attributed to the newspaper, with no confirmation from those parties in Reuters’ account.