US blocks French push for G7 coordination on frontier AI safety, calling it 'unnecessary bureaucracy'
French digital minister Anne Le Hénanff told a Senate hearing the US would not join a G7 joint statement on a high-level group to test frontier models and share incident data; a White House official cited Trump's Super Intelligence Force instead.
Why it matters
France, which holds the 2026 G7 presidency, wanted a standing group to test the most advanced models and pool incident information; without the US, the effort ends as a chair's summary rather than a joint commitment. The split comes as POLITICO notes AI systems gaining serious cyber capabilities, including a July intrusion at Hugging Face driven by an autonomous agent. Paris says it will keep going through INESIA and Inria with European partners. Watch whether the EU builds the testing group without Washington.
Key facts
- France holds the rotating G7 presidency for 2026
- Proposal: high-level working group to test frontier AI and share risk and incident information
- Le Hénanff held ministerial meetings in September and October
- US not prepared to issue a joint statement with the other six
- White House official called it 'unnecessary bureaucracy'
- France cites INESIA and Inria to continue the work