Anthropic CEO Dario Amodei and Google DeepMind CEO Demis Hassabis asked for a U.S.-led global AI coalition to help set rules and standards for artificial intelligence. They made this proposal during a closed-door lunch meeting with tech leaders and world heads at the G7 summit.
G7 summit AI discussions
The meeting took place on Wednesday at the G7 summit in Évian-les-Bains, France. The main idea from both CEOs was simple: countries should work together on AI rules, but the U.S. should lead the effort to manage risks from AI. This idea was meant to protect the world from serious AI risks, as AI systems are becoming more powerful, said a report of CNBC, citing sources who didn’t wish to be named
Canadian Prime Minister Mark Carney reportedly agreed that the U.S. could lead such an AI coalition. Companies involved in the meeting like Anthropic and Google DeepMind did not give immediate public comments after the meeting. The Office of the Prime Minister of Canada also did not respond to requests for comment at that time. The meeting came after AI systems became more advanced, especially in cyber capabilities, raising fears they could be misused for major harm.
Also read: SpaceX nears $3 trillion valuation overtaking Amazon and Microsoft, but there’s a catch
AI safety & global risks
Recently, Anthropic even disabled access to its new models called Fable 5 and Mythos 5, after U.S. export controls were placed due to national security concerns. The lunch meeting also included OpenAI CEO Sam Altman, along with around a dozen tech executives and leaders from G7 countries. The U.S. side included President Donald Trump, Treasury Secretary Scott Bessent, Commerce Secretary Howard Lutnick, and Secretary of State Marco Rubio.
Amodei said countries should cooperate on safe access to advanced AI models, and also on chip and hardware trade rules that exclude China, as cited by CNBC report, source. He also warned about AI risks in areas like cybersecurity, bioterrorism, and intelligence use, saying these need global coordination. Anthropic is still in talks with the Trump administration after the export controls on its latest AI models were introduced.
OpenAI CEO Sam Altman said there should be an international forum for AI rules, with shared testing standards and expert reviews of AI risks and abilities, as noted by a CNBC report. OpenAI also recently said its model variation GPT-5.5 Cyber is being tested in limited use for cybersecurity teams. OpenAI’s global affairs chief Chris Lehane said that even non-U.S. leaders at the meeting agreed the U.S. could take a leading role in setting AI standards.
