World leaders want American AI. They just don't want America to be able to turn it off.

World leaders at the G7 Summit are pressing for sustained access to American AI while warning against the risk that Washington could switch it off for other countries without warning. Macron and Modi warned that if the U.S. can shut off access overnight, it could disrupt European economies and undermine the AI firms themselves. The issue has intensified after the Trump administration blocked Anthropic’s Mythos 5 and Fable 5 exports on national security grounds, following an Amazon notice about guardrails that could be bypassed. During a lunch with AI executives, including Anthropic’s Dario Amodei, OpenAI’s Sam Altman, and President Donald Trump, leaders discussed a proposed 'trusted partners' framework that would let non-U.S. nations access advanced models while preserving defenses against rivals like China. Modi pressed democratic nations to secure unfettered access to leading AI models to protect critical infrastructure. Cohere CEO Aidan Gomez said the restriction underscores the danger of depending on a handful of big tech firms for digital sovereignty. Officials described a 'trusted partners' network as a potential workaround to preserve open AI access, but details on scope, eligibility, and geographic reach remained unclear. Macron supported the idea in principle, signaling Washington should consider broader Mythos access where feasible, while recognizing security concerns.