G7 weighs limited access to top US AI models for 'trusted partners'

G7 leaders meeting at a summit in Évian-les-Bains, France, are discussing a mechanism that would grant access to the most advanced AI models built by US firms such as Anthropic only to "trusted partners," it emerged on June 16. Reuters reported the talks citing several diplomatic sources. The framework would partly bypass US restrictions on non-American use, allowing allied nations and trusted organizations to use the high-end models.