G7 leaders discuss granting trusted partners access to advanced US AI models

Leaders of the Group of Seven wealthy nations are discussing a proposal to grant select "trusted partners" access to advanced AI models from U.S. firms like Anthropic, according to three diplomatic sources. This initiative could potentially bypass current restrictions on non-American use of these cutting-edge technologies.

Anthropic recently disabled access to its most advanced AI models, Fable 5 and Mythos 5, following a directive from U.S. President Donald Trump to block foreign nationals due to national security concerns. The discussion among G7 leaders, including with U.S. Commerce Secretary Howard Lutnick, aims to broaden access to these powerful AI tools.

“

trusted partners

Such an agreement would enable G7 countries to leverage these models for developing enhanced cybersecurity defenses against rivals like China. Cybersecurity experts note that Anthropic's Mythos model, designed to identify flaws in computer code, could significantly bolster attacks on banking systems, prompting interest from entities like the European Union for study.

AI executives from Anthropic, OpenAI, and Google are expected to participate in a working lunch to discuss technology issues, including regulation and AI infrastructure. The news of the "trusted partners" scheme was initially reported by the Financial Times.

“

an open line of communication with our allies, and we remain committed to addressing national security concerns with Anthropic’s model.