The Évian G7 Closed. Nine Declarations. One That Matters for AI.

Three days in Évian. Nine declarations adopted. One working lunch dedicated to artificial intelligence. President Macron held the closing press conference this afternoon. The summit is over.
Now the work of interpreting what actually happened begins. gafam.ai has been watching from the start.

The AI Working Lunch — What It Was, What It Was Not

The most significant AI moment of the entire summit was not a formal session. It was a working lunch held on the final day, June 17, attended by G7 countries, partner countries, and business leaders.

The structure tells the story. The G7 did not produce a stand-alone AI declaration. The AI conversation happened over lunch — a format that signals informality, off-the-record dialogue and conclusions that do not require unanimous communiqué language.

Three AI lab CEOs were physically present in France during these three days. Sam Altman of OpenAI. Dario Amodei of Anthropic. Demis Hassabis of Google DeepMind. The first time in history all three rival AI lab leaders appeared before world leaders simultaneously. The format of their engagement — a working lunch, not a formal session — defined what the G7 could and could not produce on AI.

The Macron Method — France's Unprecedented Approach

The most significant innovation of Évian was structural. President Macron involved partner countries — India, Kenya, South Korea, Brazil, Egypt — throughout the discussions. France also engaged civil society and the private tech sector, as this is essential for effective exchanges on issues of digital and AI regulation and innovation. This method is unprecedented.

Read that carefully. The G7 — historically an exclusive club of seven advanced economies plus the European Union — opened its discussions to five major non-G7 economies. India, in particular, was significant. Prime Minister Modi attended as an invited partner, four months after India hosted the AI Impact Summit in New Delhi. The Macron-Modi G7-India axis is now the most visible counterweight to American AI dominance among democracies.
France's method positions Europe not just as a regulator of AI, but as a convener of the global AI governance conversation. The format itself is the message. The European AI conversation now includes the Global South.

What the G7 Actually Declared

Nine declarations were adopted in Évian, covering topics from Ukraine support to economic growth and emerging technologies. The G7 leaders' joint statements include a Declaration on Mutually Beneficial International Partnerships, a Call for a Coordinated Response to the Bundibugyo Ebola Outbreak, and statements on quantum technology, critical minerals, and AI for prosperity.

What is conspicuously absent from any of these nine declarations: a binding framework for transatlantic AI export control coordination. A mutual recognition agreement for AI safety standards between the US and the EU. A specific response to the Anthropic export directive that affected European users five days ago.

The Évian G7 produced ambitious rhetoric on AI principles. It did not produce mechanisms for resolving the structural divergence between American and European AI governance that we documented throughout this week's coverage.

Trump in Évian — The Unspoken Context

US President Donald Trump addressed the media during a closing press conference at the G7 summit, defending his Iran deal as allies pressed for details.

Trump's presence at Évian was the constraint on every G7 declaration this week. The administration that killed AI safety regulation in three phone calls — the administration that ordered Anthropic to lock European users out of Claude Fable 5 last Friday — cannot simultaneously sign G7 declarations that would constrain its own AI policy.

The summit produced the strongest declarations possible given that constraint. The constraint was not French diplomacy or European political will. It was the structural incompatibility between the current US administration's AI policy and the kind of declarations that European partners would actually want to sign.

The Macron Closing Press Conference — What He Did Not Say

President Macron held the closing press conference this afternoon — a substantial address covering the full range of summit outcomes.

What we are watching for in his remarks is not what he says about AI achievements. It is what he says about AI shortcomings. A diplomatically successful closing statement would emphasise the working lunch, the partner country participation, the nine declarations. A strategically honest closing statement would acknowledge the gap between the AI conversation France hosted and the binding commitments that gap revealed.

European media will judge Macron's closing press conference on which version he delivered. gafam.ai is watching for the same signal. The difference matters for how Europe positions itself in the next 50 days — before the EU AI Act enforcement deadline arrives.

The European Perspective