G7 Nations Explore Access To Cutting-Edge US AI Systems For Trusted Allies: Report

G7 leaders have discussed a proposal to allow selected “trusted partners” access to advanced artificial intelligence models developed by U.S. companies such as Anthropic, according to three diplomatic sources. The move could create a pathway for allies to use cutting-edge AI systems despite recent restrictions on access for non-American users.

Last week, Anthropic suspended access to its most advanced AI models, Fable 5 and Mythos 5, after U.S. President Donald Trump directed the company to block foreign nationals from using its top-tier AI technologies, citing national security concerns.

One diplomatic source said representatives from several countries attending the annual G7 summit explored the possibility of expanding access to advanced AI models during discussions with U.S. officials. Much of the conversation reportedly took place with U.S. Commerce Secretary Howard Lutnick on the sidelines of Monday’s opening summit dinner in the French resort town of Evian-les-Bains.

According to a second source, the proposed “trusted partners” framework could apply to both countries and companies, though talks remain ongoing.

Supporters of the proposal argue that broader access to advanced AI systems would help G7 nations strengthen cybersecurity defenses and better counter technological challenges from rivals such as China.

In a statement, a White House official said the Trump administration maintains “an open line of communication with our allies” and remains committed to addressing national security concerns related to Anthropic’s AI models.

Executives from leading AI firms, including Anthropic, OpenAI and Google, are expected to participate in a G7 working lunch on Wednesday, where discussions are likely to focus on AI regulation, infrastructure, networks and other technology-related issues.

Anthropic did not immediately comment on the reported discussions.

Cybersecurity experts have expressed particular interest in Mythos, an AI model designed to identify vulnerabilities in computer code. While the technology could strengthen digital defenses, experts also warn it could potentially be used to enhance cyberattacks against critical systems, including those operated by financial institutions. The European Union has reportedly sought access to the model to assess its broader implications.

Before the Trump administration’s restrictions were introduced, Anthropic had provided Mythos access to selected organizations in more than 15 countries, enabling them to scan systems for security weaknesses. Those organizations operated across sectors such as healthcare, communications, energy and water infrastructure.

The Financial Times was the first to report details of the proposed “trusted partners” initiative.