G7 Leaders Negotiate "Trusted Partners" Access to Advanced US AI Models

G7 leaders are discussing a "Trusted Partners" mechanism to grant allied nations and specific companies access to advanced U.S. artificial intelligence (AI) models. The initiative aims to bypass recent U.S. restrictions on cutting-edge AI technology exports, addressing allies' national security needs. Discussions took place during the G7 summit around June 17, 2026. The proposed framework emerges as the U.S. strengthens controls on advanced AI, exemplified by Anthropic's recent halt of overseas access to its Fable 5 and Mythos 5 models, following a White House directive. These restrictions, ordered by the Trump administration, stem from concerns that advanced AI poses national security risks if unregulated. G7 members argue that denying allies access to these tools could weaken their cybersecurity and digital defense capabilities against rivals like China. Anthropic's Mythos model, designed for code analysis and vulnerability detection, is considered a dual-use tool. The discussions, involving G7 representatives and U.S. Commerce Secretary Howard Lutnick, seek to balance U.S. security interests with allied technological cooperation, potentially mirroring exemptions in semiconductor export controls.