G7 Leaders Weigh Rules for Sharing Top AI Tools with Closest Allies

G7 leaders recently met in France to talk about a new way to share America’s most advanced AI tools with a small group of trusted countries and companies. These tools, like the ones made by Anthropic, were blocked for foreign users last week because U. S. officials worry they could be used for cyberattacks or other threats. But now, some leaders want to create a special list of "trusted partners" who could still get access. Diplomats say the idea came up during a private dinner at the lakeside resort where the summit was held. The U. S. Commerce Secretary was one of the people discussing it, though no official decision was expected right away. Some leaders want this arrangement to apply to both nations and businesses, but details are still being worked out. Meanwhile, the EU is pushing to study one of these AI models, called Mythos, which could help find weaknesses in bank security systems.Cybersecurity experts warn that tools like Mythos might also help hackers find vulnerabilities in critical infrastructure. Right now, European banks aren’t getting any access, even though the EU wants to understand how these models work. The talks are happening as governments try to balance innovation with security risks. The push for a "trusted partners" system shows how hard it is to control who gets access to powerful technology. While the U. S. wants to protect its lead in AI, other countries argue they need these tools to keep up. The debate isn’t just about technology—it’s about trust and who gets to use the most advanced systems first.