AI Executives Join G7 as US and China Diverge on Tech Governance

World leaders from the G7 nations gathered in Evian, France, this week for their annual summit, but the real attention shifted to the tech executives who joined them for a private lunch. Sam Altman of OpenAI, Dario Amodei of Anthropic, and Demis Hassabis of Google DeepMind sat down with heads of state to discuss frontier AI risks, data infrastructure, and national sovereignty over emerging technologies.

A dozen other AI company leaders from Europe, Canada, India, and Japan also attended, including Arthur Mensch of France's Mistral and Aidan Gomez of Canada's Cohere. The Élysée Palace framed the session around protecting children online and preventing unchecked AI development. Yet the presence of these private sector figures underscored a deeper reality that government officials now rely on the very companies building the systems to shape any global rules.

Analysts noted the shift in influence. A senior fellow at the Council on Foreign Relations observed that credible commitments on AI require cooperation from the handful of executives actually developing the technology. This arrangement leaves elected governments negotiating alongside unelected innovators who control critical capabilities. Discussions reportedly touched on giving trusted partners limited access to advanced American AI models, a step that could further concentrate advantages among select allies while restricting broader distribution.

China responded from outside the summit. Senior officials in Beijing released a white paper promoting a global AI cooperation body open to all nations, with emphasis on serving human needs rather than creating exclusive clubs. Foreign Minister Wang Yi stressed the importance of including developing countries and avoiding trade barriers that hinder technology sharing. Chinese models have focused on affordable or freely downloadable systems, contrasting with the subscription-heavy approach common among leading Western firms.

The G7 format itself drew implicit contrast. The group of wealthy democracies has long positioned itself as a steering committee for international standards, yet its closed nature leaves major players like China to pursue parallel tracks. Beijing criticized what it called monopolistic tendencies in AI governance and signaled readiness to work directly with nations outside traditional Western alliances. This split risks fragmenting standards at a moment when AI capabilities advance rapidly across borders.

Protection of children surfaced as a shared talking point, though enforcement details remained vague amid the broader strategic conversations. Infrastructure questions also featured prominently, with questions over who controls the massive computing resources required for training advanced models. Several participants represent companies headquartered in the United States or closely aligned with American interests, raising questions about how much genuine sovereignty smaller nations can retain when core technology resides elsewhere.

The summit wrapped without any joint statement involving China, leaving two distinct approaches on display. One centers on selective partnerships among established powers and leading tech firms. The other promotes wider access with fewer gatekeepers. For observers focused on American interests, the gathering highlighted both the leverage held by domestic AI companies and the challenges of maintaining technological edges without ceding ground to competitors operating under different rules. As talks concluded, the influence of private executives at the table stood as a clear marker of how power has redistributed in this domain.