G7 Explores Special Access to US AI Models for Trusted Partners

June 16, 2026

Mintesinot Nigussie

G7 officials have discussed a possible arrangement that would allow selected countries or companies to access advanced US artificial intelligence models despite new restrictions on foreign access, according to three diplomatic sources cited by Reuters.

The talks focused on creating a group of “trusted partners” that could receive exemptions from Washington’s recent decision to limit access to Anthropic’s most advanced AI systems for non-US nationals.

The proposal was discussed with US representatives, including Commerce Secretary Howard Lutnick, during the opening G7 summit dinner in Evian-les-Bains, France, one of the sources said.

The idea remains under discussion, and potential beneficiaries could include either governments or companies, another source said. No final decision or public announcement was expected during the G7 meeting’s technology discussions on Wednesday, a third source said.

Washington introduced the restrictions on Anthropic’s frontier AI models last week, citing national security concerns. The move has increased debate among governments over how access to advanced AI technologies should be managed internationally.

The discussions at the G7 come as major economies seek ways to balance technological cooperation with concerns over security, competitiveness and control of high-end AI capabilities.