CEOs of Anthropic, Google, and DeepMind Urge US-Led AI Coalition at G7

In a pivotal moment for the future of artificial intelligence governance, top executives from leading AI firms, including Anthropic and Google DeepMind, have called for the establishment of a U.S.-led international coalition to develop robust rules and standards for the rapidly evolving technology. This significant proposal emerged from a closed-door working lunch held on Wednesday at the G7 summit in Évian-les-Bains, France, a gathering that brought together global tech leaders and heads of state.

Dario Amodei, CEO of Anthropic, and Demis Hassabis, CEO of Google DeepMind, both emphasized the urgent need for coordinated international efforts, with the United States spearheading the initiative, to mitigate the potential risks associated with advanced AI. Sources close to the discussions, who requested anonymity due to not being authorized to speak publicly, revealed the details of the private session.

Canadian Prime Minister Mark Carney reportedly expressed agreement on the U.S. taking a leading role in such an AI coalition, according to one individual familiar with the deliberations. While Anthropic has declined to comment on the G7 meeting, and Google DeepMind and the Office of the Prime Minister of Canada did not immediately respond to requests for comment, the sentiment expressed at the summit underscores a growing consensus among industry titans and policymakers regarding AI’s profound implications.

This high-level discussion comes at a critical juncture, following the release of increasingly sophisticated AI models with advanced cyber capabilities. Concerns have mounted within the industry regarding the potential for these powerful tools to be misused, leading to significant disruptions or even major disasters if they fall into the wrong hands. Anthropic’s recent decision to disable access to its cutting-edge models, Fable 5 and Mythos 5, on Friday, in response to U.S. government export controls citing national security concerns, highlights the immediate challenges and the proactive measures being considered by both companies and governments.

The meeting saw the participation of approximately a dozen tech executives, including OpenAI CEO Sam Altman, alongside leaders from the G7 nations. Representing the United States at the summit were Secretary of the Treasury Scott Bessent, Secretary of Commerce Howard Lutnik, and Secretary of State Marco Rubio, alongside President Trump.

During his address, Amodei proposed that international cooperation should encompass structured access to frontier AI models and the controlled trade of essential components like chips, explicitly excluding China from such supply chains. He further advocated for collaborative efforts among nations to address critical risks posed by AI in areas such as cybersecurity, bioterrorism, and intelligence operations. Anthropic is currently engaged in ongoing negotiations with the U.S. administration following the implementation of export controls on its latest models.

OpenAI’s CEO, Sam Altman, echoed the sentiment for international collaboration, calling 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.” This aligns with OpenAI’s recent proactive measures, including the limited preview rollout of GPT-5.5 Cyber to vetted cybersecurity teams last month, a specialized version of its latest model designed to address sophisticated cyber threats.

Chris Lehane, Global Head of Policy at OpenAI, who was also present at the G7 meeting, noted that leaders from countries outside the U.S. acknowledged the potential for the United States to “certainly play the lead role in working to establish” globally recognized AI standards. This suggests a willingness among international partners to support U.S. leadership in shaping the future of AI governance.