G7 leaders are discussing a proposal to provide select “trusted partners” access to cutting-edge artificial intelligence models from U.S. firms, according to multiple diplomatic sources. The discussions center on allowing access through a controlled arrangement that could broaden G7 countries’ ability to use advanced AI systems developed in the United States, potentially creating a pathway around existing restrictions on using non-American-developed models.
The plan is described as involving access from U.S. AI companies such as Anthropic, with the term “trusted partners” referring to countries or entities that meet specified criteria. The proposed use case highlighted by sources is cybersecurity: G7 members would use the advanced models to strengthen defenses and improve capabilities against cyber threats.
All accounts also note the geopolitical context for the cybersecurity focus, including concerns about rivals such as China. The reporting indicates the discussions are at the source level and does not describe final approval, timelines, or the full terms of any agreement.
