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

World leaders gathered in Evian, France, this week found themselves sharing the spotlight with a roster of Silicon Valley and tech executives at the G7 summit, underscoring how a handful of private companies now shape discussions once reserved for governments. CEOs from OpenAI, Anthropic, Google DeepMind and other firms joined sessions focused on frontier AI risks, infrastructure needs and digital sovereignty, while protection of children online featured prominently on the agenda.

The presence of these executives reflects a clear transfer of influence. As one analyst noted, credible international commitments on artificial intelligence now require the cooperation of the small group of firms actually developing the most advanced systems. That reality was on display during a lunch meeting that brought together figures such as Sam Altman, Dario Amodei and Demis Hassabis alongside leaders from Mistral, Cohere and several other companies. Their input is being sought on issues that will affect regulation, access and security for years to come.

At the same time, China moved to present an alternative vision. Senior officials in Beijing released a white paper on global AI governance that called for an inclusive international body open to all countries and criticized closed, monopolistic approaches to the technology. They stressed the need to serve human needs broadly and signaled willingness to share models more freely, often at low or no cost, in contrast to subscription-based Western offerings. Chinese diplomats framed their push as support for developing nations that risk being left behind by restrictive export controls and selective partnerships.

The G7 discussions reportedly included plans to grant trusted partners access to leading U.S. models while maintaining tight limits on others. Such steps risk deepening divisions at a moment when AI capabilities are advancing rapidly across borders. Critics argue that framing access around geopolitical alignment rather than shared safety standards could slow collective efforts to address genuine risks, from misuse by bad actors to uncontrolled deployment.

France, as host, highlighted child safety as a priority, yet the broader conversation remains dominated by the same firms whose products have already raised concerns about data practices and algorithmic amplification of harmful content. The summit format itself illustrates how traditional multilateral forums are adapting to corporate power rather than constraining it.

Beijing’s outreach, by comparison, positions China as an advocate for wider participation, particularly from the Global South. Whether that stance translates into concrete safeguards remains to be seen, but the contrast with G7 messaging is deliberate. One side emphasizes selective partnerships and controlled diffusion; the other promotes open cooperation while accusing rivals of exclusionary tactics.

The outcome leaves unresolved questions about accountability. When unelected executives help set the terms for global rules on a transformative technology, democratic oversight becomes harder to maintain. Meanwhile, the absence of China from core G7 talks limits the scope of any agreements reached. As AI systems grow more capable, the summit’s emphasis on trusted circles may prove less durable than claims of inclusive governance.