Full Breakdown
G7 Leaders and AI CEOs Tackle U.S. Export Controls and a “Trusted Partners” Scheme
6/17/2026, 9:46:47 PM
The AI Agenda at the G7 Summit
From June 15-17, the Group of Seven gathered in Evian-les-Bains, France, to address a range of security issues. On the final day, heads of state met with senior executives from the world’s leading artificial-intelligence firms for a working lunch focused on “ensuring a safe, rapid and effective deployment of artificial intelligence.” The agenda included discussion of a proposed “trusted partners” mechanism that could grant limited access to U.S. frontier AI models for selected allies.
Background: U.S. Export Controls on Anthropic Models
On June 16, the Trump administration ordered Anthropic to block foreign nationals from its most advanced models, Fable 5 and Mythos 5, citing national-security concerns. Anthropic complied by disabling global access, noting that the models had previously been available in “more than 15 countries.” The move sparked immediate debate among G7 participants about the geopolitical implications of restricting advanced AI technology.
Key Participants
- AI CEOs: Dario Amodei (Anthropic), Sam Altman (OpenAI), Demis Hassabis (Google DeepMind), Arthur Mensch (Mistral AI), Marc Benioff (Salesforce), Alex Wang (Meta).
- U.S. officials: Commerce Secretary Howard Lutnick, senior White House officials overseeing the export-control order.
- European officials: President Emmanuel Macron, European Commission President Ursula von der Leyen, Commission spokesperson Thomas Regnier, EU tech chief Henna Virkkunen, German Chancellor Friedrich Merz.
Proposed “Trusted Partners” Scheme
Diplomatic sources reported that G7 delegates explored a framework allowing “trusted partners”—either sovereign states or companies—to obtain exemptions from the U.S. ban on Anthropic’s frontier models. The scheme is intended to enable allies to use the models for cybersecurity research, particularly against threats from China, while preserving U.S. national-security safeguards. No formal announcement was expected at the summit.
Divergent Views on AI Sovereignty
European officials expressed concern that unilateral U.S. restrictions could create a “kill-switch” capability, undermining the reliability of critical AI tools. The European Commission emphasized the need for a “circle of trust” and warned against discriminatory access that could disadvantage partners. Simultaneously, U.S. officials framed the restrictions as a necessary response to a “potential jailbreak” that might allow the models to identify software vulnerabilities.
Official Statements & Responses
President Macron said progress on broader access to U.S. models was likely in the coming weeks and framed cooperation as mutually beneficial. Thomas Regnier described the EU’s stance as readiness to “engage and tackle these security risks together with our like-minded partners.” Ursula von der Leyen highlighted the interdependence of U.S. and EU financial systems and called for shared use of trusted technology.
Verbatim Quotes
- “We need to establish American AI [...] as the gold standard for AI worldwide and ensure our allies are building on American technology,” — Trump administration AI Action Plan, July 2025
- “We are ready to engage and tackle these security risks together with our like-minded partners,” — Thomas Regnier, European Commission spokesperson
- “We develop the technology, and the citizens of the free world make the rules.” — Sam Altman, OpenAI CEO
- “The Anthropic kill switch shows that tech sovereignty was never abstract,” — Brando Benifei, MEP (S&D, Italy)
- “We use each other’s trusted technology, and our financial systems are interconnected,” — Ursula von der Leyen, European Commission President
Impact and Outlook
The discussion signaled a potential shift toward more coordinated transatlantic AI governance, while also reinforcing European calls for greater technological autonomy. If the “trusted partners” scheme materializes, it could reshape access to frontier models and influence future cybersecurity collaborations.
What’s Next
G7 officials plan to task finance ministries, regulators, and cybersecurity experts with assessing the economic and security implications of frontier AI. The European Commission is advancing its AI “gigafactory” and sovereign-cloud initiatives, and further negotiations on the trusted-partner framework are expected in the weeks ahead.
