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

World leaders gathered in Evian, France, this week for the annual G7 meetings found themselves sharing space not only with traditional diplomatic counterparts but also with a group of prominent artificial intelligence executives. The presence of figures such as OpenAI chief Sam Altman, Anthropic CEO Dario Amodei and Google DeepMind head Demis Hassabis underscored how decisions about advanced technology now require direct input from the companies developing it.

The gathering focused on frontier AI risks, computing infrastructure and questions of national sovereignty over digital systems. French officials also highlighted child online safety as a priority topic. Alongside the major U.S. and European firms, executives from Mistral, Cohere, Synthesia and several Asian AI startups joined a lunch session with heads of state. Observers noted the unusual format as evidence that governments can no longer set credible rules for the technology without cooperation from its leading developers.

China, absent from the closed-door sessions, used the same week to release its own blueprint for global AI governance. Senior diplomat Wang Yi described plans for an inclusive international AI body open to all countries, with particular outreach to developing nations. Chinese officials contrasted their approach with what they called closed and monopolistic models, pointing to efforts to distribute capable systems at low or no cost rather than through paid subscriptions.

The contrast reflects deepening divisions over who should control access to advanced models. Reports from the summit indicated G7 members are exploring ways to route U.S.-origin systems toward trusted partners while limiting broader diffusion. Beijing has framed its own offerings as an alternative route that avoids such restrictions, pairing the technical push with criticism of trade barriers that it says hinder equitable progress.

Policy analysts tracking the meetings described the inclusion of private executives as a structural shift rather than a temporary arrangement. Because the most capable systems are concentrated in a handful of firms, any agreement on safety standards, export controls or infrastructure investment now depends on aligning government objectives with corporate incentives. That reality has altered the composition of summits that once centered almost exclusively on elected officials and career diplomats.

At the same time, the exclusion of China from these conversations has accelerated separate tracks of rule-making. Beijing's white paper emphasized support for the Global South and rejected what it portrayed as exclusionary frameworks. The result is two parallel conversations about AI governance, each seeking to set norms that the other side views as insufficient or self-serving.

The coming months will test whether these overlapping efforts can produce compatible standards on issues such as model evaluation and security protocols. For now, the Evian meetings illustrated how power over artificial intelligence is distributed across both state capitals and corporate boardrooms, with neither side able to dictate outcomes alone.