G7 Leaders Sound Alarm Over U.S. AI Export Vulnerability

At the G7 Summit on Wednesday, world leaders including French President Emmanuel Macron and Indian Prime Minister Narendra Modi voiced deepening concerns that the United States could unilaterally cut off their countries’ access to leading American artificial intelligence models at any moment. The warning came during a working lunch that brought together top AI executives such as Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman, alongside President Donald Trump.

Macron’s Cautionary Tale

Macron cautioned that if the U.S. “from one day to the next can turn off the switch,” the consequences would ripple far beyond European customers. He argued that such abrupt disruption would not only harm the economies of allied nations but also damage the AI firms themselves by eroding trust in their reliability. The French president’s remarks reflected a broader unease among non-U.S. nations about the fragility of their dependence on American-controlled AI infrastructure.

The Anthropic Blockade

The controversy erupted days earlier when the Trump administration blocked Anthropic from exporting its newest large language models — Mythos 5 and Fable 5 — on national security grounds. The ban followed Amazon’s flagging to the White House that certain safety guardrails in the models could be bypassed. Despite cybersecurity experts arguing that similar capabilities exist in freely available models from competitors such as OpenAI, Anthropic’s exports remain frozen.

This episode has laid bare a risk that many international companies have been grappling with: any enterprise or government that builds on U.S. AI infrastructure now has to reckon with the possibility that access can be revoked overnight, for reasons that may never be fully disclosed. The uncertainty creates a chilling effect on innovation and investment in nations that rely on cutting-edge American models.

Modi’s Call for Unfettered Access

Prime Minister Modi also expressed alarm at Trump’s move. According to reporting from the Financial Times, Modi argued that democratic nations must have unfettered access to top-tier AI models to protect critical infrastructure — from power grids to healthcare systems — from hostile actors. He stressed that restrictions on AI exports could undermine collective security and economic resilience.

“The recent restriction on access to Anthropic’s models confirms what we at Cohere have known all along: that companies and democratic nations remaining dependent on a small handful of big tech companies is dangerous to resilience,” said Aidan Gomez, co-founder and CEO of Canadian enterprise AI firm Cohere, in a statement shared with TechCrunch. “Digital sovereignty is not just about market competition or any one company or nation. It’s about who controls the foundational technology that will shape our economic security and national sovereignty for decades to come.”

Trusted Partners Scheme

During the meeting, G7 leaders explored the creation of a “trusted partners” scheme that would grant non-U.S. nations access to advanced AI models from firms like Anthropic and OpenAI without going through standard U.S. export controls. The goal is to maintain an open trade network that could bypass potential restrictions. Both countries and companies could qualify as trusted partners, provided they use the models to develop stronger defenses against rivals like China.

However, the specifics of the scheme remain hazy. It is unclear how far the trusted partner designation would extend or whether it offers any relief for a startup in Paris or Bangalore that suddenly finds its product broken because access to a foundational model was cut. Macron noted that it would make strategic sense for Washington to back such a scheme and ensure wider access to Mythos. He argued that nobody would want to buy U.S. AI access if it could disappear overnight.

The Sovereignty Dilemma

The comments come amid broader efforts by Europe and other non-U.S. countries to push for AI sovereignty. Yet building independent AI capabilities is an increasingly difficult proposition when American models continue to pull ahead in performance benchmarks. Nations like France, Germany, and India have invested heavily in domestic AI research, but they still rely on U.S. platforms for the most advanced capabilities. The Anthropic block demonstrates that even temporary restrictions can have cascading effects on supply chains, product development, and national security.

“The recent restriction on access to Anthropic’s models confirms what we at Cohere have known all along: that companies and democratic nations remaining dependent on a small handful of big tech companies is dangerous to resilience,” Aidan Gomez reiterated, emphasizing that control over foundational AI technology is a matter of economic and national security.

Broader Implications for Global AI Governance

The G7 summit’s focus on AI export controls highlights a fundamental tension in global technology governance. On one hand, the United States seeks to prevent powerful AI capabilities from falling into the hands of adversaries. On the other, allies demand predictability and trust in the systems they integrate into their economies. Without clear rules and guarantees, the current patchwork of export bans and licensing requirements risks fracturing the international AI ecosystem.

Some experts argue that the solution lies in multilateral agreements that set shared standards for AI safety and export controls. A transparent framework could reassure allies while still addressing national security concerns. But reaching consensus among the G7, let alone the broader international community, is a daunting task. The Anthropic case underscores how quickly unilateral actions can undermine collective efforts.

Meanwhile, companies like Anthropic and OpenAI face pressure to balance innovation with responsibility. The fact that Amazon flagged potential vulnerabilities in Anthropic’s models suggests that safety guardrails are a genuine concern. But the opaqueness of the government’s decision-making process fuels suspicion. Without clear criteria for what triggers a block, the entire industry operates under a cloud of uncertainty.

Voices from the Industry

Industry leaders have weighed in on the controversy. While some support robust export controls to prevent misuse, others warn that excessive restrictions could push allied nations to develop their own AI ecosystems, potentially fragmenting the market and slowing global progress. The tension is especially acute in Europe, where regulators are already charting a different path through the AI Act and national initiatives.

In India, the government has accelerated its own AI mission, aiming to build sovereign capabilities in language models and data infrastructure. Modi’s concerns at the G7 reflect a broader desire for strategic autonomy in a domain that increasingly defines economic and military power.

As the G7 summit continued, discussions turned to how to structure the trusted partners scheme. Macron suggested that it could include explicit commitments to reciprocity and transparency. No final decisions were announced, but the fact that the issue dominated the agenda signals a major shift in the geopolitics of AI.

In the end, the episode serves as a wake-up call: the age of effortless access to American AI may be coming to an end. Nations and companies alike must prepare for a landscape where access is a privilege that can be revoked, not a right. The question now is whether the United States and its allies can build a new framework that balances security with openness — or whether the world will splinter into competing AI blocs.

Source: TechCrunch News