G7 leaders on Wednesday pledged closer coordination on the risks and opportunities of advanced AI, while also discussing the creation of a "trusted partners" scheme granting access for non-U.S. ​nations to advanced U.S. AI models from firms like Anthropic.