G7 Évian opens Monday with all three AI CEOs on site — the first summit where frontier AI has a seat at every table

The 52nd G7 summit opens Monday in Évian-les-Bains, France, and for the first time all three frontier AI lab CEOs are confirmed: OpenAI's Sam Altman (personally invited by President Macron), Anthropic's Dario Amodei, and Google DeepMind's Demis Hassabis. The trio co-signed a letter to Congress on AI-related biosecurity risks earlier this month — a rare alignment that sets an unusual tone going in.

What's on the agenda. AI is a primary focus. Youth safety sits at the top of OpenAI's agenda: Altman is proposing a Global Youth AI Safety Institute to coordinate age-appropriate standards and annual risk assessments across member nations. Frontier AI risks in the cyber and biological domains are the other major track, directly connected to the biosecurity letter.

US positioning. The Trump administration has made clear it opposes new multilateral agreements that could constrain US competitive advantage in AI, per TechPolicy.Press. Expect the final communiqué to lean heavily on voluntary commitments rather than binding agreements — "responsible use" language rather than mandatory pre-release certification.

Why it matters to builders. Amodei's June 10 essay calling for FAA-style pre-release safety certification lands as formal negotiating context for these discussions. If G7 nations move even partially toward coordinated frontier model oversight, the compliance burden ultimately hits the teams shipping those models and the developers who integrate them. This weekend won't produce binding law — but it can set the framing that does.

Watch for AI language in the G7 communiqués published after June 17. Any reference to coordinated "frontier model" evaluation or reporting requirements is worth reading closely.

Sources: CNBC: France invites Altman to G7, Dataconomy: AI leaders at G7, The Next Web: AI rivals head to G7, TechPolicy.Press: rift over AI sovereignty, OpenAI: youth AI safety proposal