Anthropic and DeepMind CEOs Push for U.S.-Led AI Coalition at G7
Dario Amodei and Demis Hassabis proposed international cooperation on AI standards and security at a closed-door meeting with tech leaders and heads of state in France.
The chief executives of two leading AI companies called for the United States to lead an international coalition on artificial intelligence governance during a private meeting with world leaders at the G7 summit in France on Wednesday.
Dario Amodei of Anthropic and Demis Hassabis of Google DeepMind both advocated for U.S.-led international cooperation to establish rules and standards around AI technology, according to two people with knowledge of the discussions. Canadian Prime Minister Mark Carney expressed support for the U.S. taking a leadership role in such a coalition.
The meeting in Évian-les-Bains brought together approximately a dozen technology executives, including OpenAI CEO Sam Altman, with G7 heads of state. President Donald Trump attended alongside Treasury Secretary Scott Bessent, Commerce Secretary Howard Lutnik, and Secretary of State Marco Rubio.
Proposed areas of cooperation
Amodei outlined specific domains where international collaboration should focus, according to sources familiar with his remarks. These included structured access protocols for frontier AI models, coordination on semiconductor and critical component trade that excludes China, and joint efforts to address AI-related risks in cybersecurity, bioterrorism, and intelligence operations.
Altman echoed the call for coordination, advocating for "an international forum for discussion that establishes globally accepted standards for testing, provides expert and impartial analysis of capabilities and risks, and serves as a venue for cooperation among nations," according to an OpenAI briefing.
Chris Lehane, OpenAI's global affairs chief who attended the meeting, said non-U.S. leaders acknowledged that the United States "certainly could play the lead role in working to establish" AI standards.
Why it matters
The push for international AI governance comes as advanced models demonstrate capabilities that have alarmed security experts. The U.S. government imposed export controls on Anthropic's newest models, Fable 5 and Mythos 5, late Friday, citing national security concerns. Anthropic disabled access to those models and remains in negotiations with the Trump administration. The episode illustrates the tension between AI companies' global ambitions and governments' security imperatives—a dynamic that a formal coalition structure might help manage. With China excluded from proposed chip-trade arrangements, the meeting signals an emerging technology bloc strategy reminiscent of Cold War-era alliances.
The timing reflects growing unease about AI systems' potential for misuse. OpenAI recently released GPT-5.5 Cyber, a specialized variant of its latest model, in limited preview to vetted cybersecurity teams—an acknowledgment that the same capabilities useful for defense can enable sophisticated attacks.
These details were first reported by CNBC.
This is an original analysis by the Omega editorial team. Source reporting: AI Watch.
Want systems like this working for your business?
Book a Call
