G7 Leaders Push Back Against U.S. Control Over AI Access

The Export Control Flashpoint

According to TechCrunch, the Trump administration recently blocked Anthropic from exporting its newest models on national security grounds. The action followed a report from Amazon to the White House that certain safety guardrails in the models could be circumvented. Despite cybersecurity experts noting that comparable capabilities exist in freely available models from competitors like OpenAI, Anthropic’s models remain restricted from export.

This move crystallized a recurring concern among international leaders: U.S. companies and governments cannot reliably depend on long-term access to American AI infrastructure.

Macron and Modi Voice Alarm at G7 Summit

At the G7 Summit on June 17, French President Emmanuel Macron warned fellow leaders and AI executives—including Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, and President Donald Trump—that allowing the U.S. to unilaterally cut off access to AI models poses a systemic risk. According to TechCrunch, Macron argued the threat of overnight access revocation would undermine both customer economies and the profitability of AI firms themselves.

Indian Prime Minister Narendra Modi echoed similar concerns, according to reporting from the Financial Times cited by TechCrunch. Modi emphasized that democratic nations require unfettered access to advanced AI models to safeguard critical infrastructure. The convergence of concern across major economies signals that the Anthropic export block has moved export control from a theoretical risk to an observable pattern.

The Digital Sovereignty Response

The conversation at the G7 has surfaced a broader anxiety about technological dependency. TechCrunch notes that during the summit, leaders discussed creating a “trusted partners” scheme—a framework intended to grant non-U.S. nations reliable access to advanced AI models from firms like Anthropic and OpenAI. The mechanism would operate as an alternative trade network that circumvents unilateral U.S. restrictions, with membership tied to commitment to developing defenses against geopolitical rivals like China.

Aidan Gomez, co-founder and CEO of Canadian enterprise AI firm Cohere, framed the issue in terms of resilience. According to TechCrunch, Gomez stated that the Anthropic restrictions underscored why dependence on a small number of U.S. technology companies threatens national and economic security, and that digital sovereignty matters for who controls the foundational technology shaping decades of policy.

Why This Matters

The G7 discussion exposes a crack in the post-Cold War technology order: U.S. companies export AI globally, but the U.S. government retains the unilateral right to revoke access at any time, for reasons that may never be disclosed. This creates acute risk for startups in Europe and South Asia that have built products atop U.S. models—they face the possibility of sudden product failure with no recourse.

Macron’s point is economically subtle: no rational buyer commits to a supply chain they cannot trust. If the U.S. cannot credibly guarantee stable access, international adoption of American AI models will stall, or shift to open-weights alternatives from non-U.S. vendors. The “trusted partners” proposal, if formalized, would represent an attempt to ring-fence certain allies from export controls, though TechCrunch notes the framework remains vague and untested. Whether such a scheme would satisfy smaller nations and companies—or merely entrench reliance on a new set of gatekeepers—remains unresolved.