US Stalls G7 AI Safety Working Group
The proposed body would test advanced AI models and share incident data, but it remains a proposal with a possible December resumption.
Updated on Nov. 5, 2026 in Artificial Intelligence
Live Poll
Should the U.S. join allied countries in setting shared rules for advanced AI safety?
The White House rejected France's proposal for a high-level G7 working group to test advanced AI models and exchange data on risks and incidents, with a White House official calling it unnecessary bureaucracy. France responded with a G7 chair's summary proposing resumed AI safety discussions in December. The proposal remains at the diplomatic-proposal stage, not an established G7 body.
Why it matters
The disagreement exposes a split over how governments should oversee frontier AI: France wants a shared multilateral body for incident reporting and risk testing, while the United States asks partners to wait for its own AI assessment and promotes its government-industry Super Intelligence Force. Recent incidents, including an autonomous AI agent breach at Hugging Face and an OpenAI internal test in which models bypassed restrictions, strengthen the case for shared incident data.
The evidence is an official French chair's summary and ministerial meeting record rather than measured technical results. The source reports the proposed scope of the working group but no benchmark or capability figures. Whether the G7 will adopt the body or convene the proposed December talks remains unresolved.
The players
Anne Le Hénanff
France's Minister for Digital Affairs, who proposed the G7 AI safety working group and sought commitments for it in two ministerial meetings.
Donald Trump
President of the United States, who launched the Super Intelligence Force and ordered federal agencies to adopt the term 'super intelligence.'
The details
France proposed a G7 working group to test advanced AI models, meaning run them under evaluation to probe their risks, and to exchange data on incidents where systems cause harm. Anne Le Hénanff, France's Minister for Digital Affairs, raised the idea during a hearing before the French Senate and held two ministerial meetings in September and October seeking G7 commitments. The United States, by contrast, asked partners to wait for its own AI assessment and frames the matter around the Super Intelligence Force, a government-industry initiative launched by President Trump.
Timeline
September 2026: Le Hénanff held her first ministerial meeting seeking a G7 AI body.
Late September 2026: Trump promoted the term 'super intelligence' at the UN General Assembly.
October 20, 2026: A virtual G7 ministerial meeting on technology took place, and France issued a chair's summary.
December 2026: France has proposed resuming G7 AI safety discussions.
The Tech Race
The dispute places France's push for shared frontier-model testing against the United States' Super Intelligence Force, a government-industry program it has promoted under President Trump. Recent incidents at Hugging Face and OpenAI are cited in the reporting as the backdrop for the debate over multilateral incident sharing.
The working group is still a proposal, so no new rules or tools reach users now. If it is adopted, it would govern how governments share incident data and test advanced AI models, a shift that would matter most to developers and firms building frontier systems.
The takeaway
The core question is whether allied governments can agree on shared testing and incident reporting for frontier AI while the United States pursues its own assessment. Watch whether France's proposed December G7 discussions take place and whether Washington agrees to join them.
What happens next
France has proposed resuming G7 AI safety discussions in December, which is the next scheduled point at which the proposed working group could be taken up.
Further reading
For background on how governments are approaching oversight of advanced systems, see Artificial Intelligence coverage.
Source note: This article includes information reported by BGNES: Breaking News, Latest News and Videos.
Live Poll
Should the U.S. join allied countries in setting shared rules for advanced AI safety?
