World leaders crave American AI but reject its kill switch

During Wednesday's G7 Summit, world leaders including French President Emmanuel Macron and Indian Prime Minister Narendra Modi expressed concerns that the U.S. might abruptly restrict their countries' access to leading American AI models at any time.

Macron cautioned G7 leaders and prominent AI executives—among them Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, and President Donald Trump—during lunch that if the U.S. could "turn off the switch from one day to the next," it would harm not only European customers' economies but also the AI companies themselves.

These remarks came days after the Trump administration blocked Anthropic from exporting its latest Mythos 5 and Fable 5 models, citing national security. The directive followed Amazon alerting the White House that certain safety guardrails could be circumvented. Although cybersecurity experts contend that the capabilities cited by the government exist in other freely available models, including those from OpenAI, Anthropic's models remain frozen.

This incident has highlighted a risk many international firms have been wrestling with: any organization or government that relies on U.S. AI infrastructure must now face the possibility that access could be withdrawn overnight, for reasons they may never learn.

Prime Minister Modi also voiced concern over Trump's decision to block Anthropic's model, according to Financial Times reporting, adding that democratic nations need unrestricted access to top AI models to safeguard critical infrastructure.

"The recent restriction on Anthropic's models confirms what we at Cohere have long understood: that companies and democratic nations continuing to rely on a small number of big tech firms poses a danger to resilience," said Aidan Gomez, co-founder and CEO of Canadian enterprise AI company Cohere, in a statement provided to TechCrunch. "Digital sovereignty isn't just about market competition or any single company or nation. It's about who controls the foundational technology that will shape our economic security and national sovereignty for decades to come."

During the meeting, G7 leaders also explored establishing a "trusted partners" framework that would allow non-U.S. nations to access advanced AI models from companies like Anthropic and OpenAI. The aim is to preserve an open trade network that circumvents U.S. restrictions. Both countries and corporations could become trusted partners, provided they use the models to strengthen defenses against competitors such as China.

However, it remains unclear how extensive that trusted partner scheme would be, or whether it offers a solution for a startup in Paris or Bangalore whose product suddenly fails without notice.

Nonetheless, Macron noted that it would be logical for Washington to support such a scheme and to ensure broader access to Mythos. No one would want to purchase U.S. AI access if it could vanish overnight.

These comments came as Europe and other non-U.S. nations strive for AI sovereignty—an increasingly challenging argument to make as American models continue to advance and no one wants to be left behind.