US Anthropic curbs unsettle G7 AI talks

Western allies arrived at the G7 AI summit in France rattled by Washington’s decision to cut off non-Americans from Anthropic’s most powerful models, including the company’s own foreign employees. The Trump administration imposed export control measures blocking all access to Fable 5 and Mythos 5 by any foreign national, citing national security concerns tied to the models’ ability to identify and exploit cyber vulnerabilities.

The move has reinforced European fears that US technology could be withdrawn suddenly, even from trusted partners. Brussels said safeguards for high-risk AI may be justified but should not discriminate against allies, with European Commission spokesperson Thomas Regnier arguing that Europe is a trusted partner. The restrictions landed as Brussels was preparing to join the Pax Silica, a US-led alliance focused on AI chip and critical minerals supply chains.

France had made AI central to its G7 presidency, aiming to advance safe and rapid deployment while positioning Paris as a research and investment hub. The Anthropic restrictions are now expected to shadow the summit’s AI agenda and strengthen Europe’s push to reduce reliance on foreign providers and build technological sovereignty from a position of strength.