Global Leaders Confront US Over AI Access Control Amid Dependency Fears
World leaders at the G7 summit are pushing back against U.S. control over frontier AI models, following the Trump administration's abrupt export ban on Anthropic’s latest technology.
At the G7 summit, international leaders signaled growing unease over the United States' ability to unilaterally restrict access to frontier artificial intelligence. French President Emmanuel Macron and Indian Prime Minister Narendra Modi voiced concerns that the U.S. could effectively turn off the switch on critical AI infrastructure, according to TechCrunch. This apprehension follows a recent move by the Trump administration to block Anthropic from exporting its newest Mythos 5 and Fable 5 models. The administration’s decision, which forced an immediate halt to international access, has transformed AI from a commercial product into a flashpoint for national security and sovereignty discussions.
The export ban was triggered after Amazon alerted the White House that Anthropic’s latest models contained safety guardrails that could be bypassed, according to The Verge. While the administration cited national security as the justification for the restrictions, the move has faced skepticism. Cybersecurity experts noted that the capabilities in question are also present in models from other providers, such as OpenAI, which remain freely available. Reports on the timeline of the decision remain conflicting, with some accounts suggesting the White House gave Anthropic as little as 90 minutes to disable access, while other officials claimed they had negotiated for hours.
Beyond the technical justifications, the conflict highlights the fraught relationship between the current administration and certain AI firms. According to The Verge, some observers view the crackdown as a consequence of Anthropic’s perceived political opposition to the Trump administration. Regardless of the underlying motivations, the incident has served as a stark warning for international businesses and governments that rely on U.S.-based AI infrastructure. The reality, as noted by former White House AI adviser Dean Ball, is that frontier AI companies must now operate with extreme regulatory caution, as they effectively require an explicit government green light to function.
The incident has accelerated calls for technological sovereignty among non-U.S. nations. Aidan Gomez, CEO of the Canadian enterprise AI firm Cohere, stated that the reliance on a small group of big tech companies is a risk to national resilience. During the G7 meeting, leaders discussed the potential for a trusted partners scheme, which would allow non-U.S. countries to access advanced models from companies like Anthropic and OpenAI. The proposed framework aims to maintain an open trade network for AI, provided the technology is used to develop defenses against rivals like China, according to TechCrunch.
The viability of such a scheme remains uncertain, as does the broader question of how international startups can operate without the threat of sudden service disruptions. As TechCrunch reported, Macron suggested that Washington should support a broader access model, noting that international customers are unlikely to invest in U.S. AI if their access can disappear overnight. For now, the episode underscores a growing divide: while American models continue to lead in performance, the risk of centralized control is prompting foreign governments to reconsider their dependency on the U.S. AI ecosystem.