US Imposes Unprecedented AI Export Ban

In a move that has intensified tensions between the United States and its allies, the Trump administration has directed the technology firm Anthropic to cease providing foreign access to its sophisticated AI models, Mythos 5 and Claude Fable 5. Citing national security imperatives, this directive marks an unprecedented intervention in the artificial intelligence industry. The US government's order, which applies to all foreign entities both within and outside the United States, led Anthropic to take both AI models offline to ensure full compliance.

Prior to this ban, Anthropic had extended access to its Claude Mythos Preview, a cutting-edge model, to approximately 200 institutions across 15 countries for vulnerability testing. The public iterations, Mythos 5 and Fable 5, were slated for release in early June. While Anthropic stated that the US government did not articulate specific reasons for the order, the company understood that the Trump administration believed a method for 'jailbreaking' Fable 5 had been discovered.

Allies React to Restrictive Measures

The immediate aftermath of the ban saw widespread apprehension across Europe, a region heavily reliant on US-developed artificial intelligence. French President Emmanuel Macron, addressing the Group of Seven (G7) nations, acknowledged the order as a 'wake-up call' regarding AI risks, but simultaneously criticized the restrictions as detrimental. He characterized the reaction as 'strictly nationalist,' highlighting the departure from previous policies where US tech restrictions primarily targeted adversaries like China and Russia. This latest order, however, applies equally to allied nations with whom Washington shares intelligence and mutual defense agreements.

This decision is not an isolated incident but follows a pattern of transactional policy shifts by the Trump administration. Over the past 18 months, the administration has initiated a global trade war, threatened to annex Greenland (an autonomous Danish territory), and contemplated withdrawing from the 77-year-old North Atlantic Treaty Organization (NATO). Furthermore, threats were made to halt weapon supplies to Ukraine unless European allies assisted in reopening the Strait of Hormuz, which Iran had effectively closed following the US and Israeli conflict with the country on February 28.

Calls for Technological Self-Reliance

Dex Hunter-Torricke, president of the Center for Tomorrow, observed that US allies are increasingly realizing their significant vulnerability to the 'US techno-industrial complex.' President Macron underscored the necessity for international collaboration on AI matters, cautioning against the perils of 'non-cooperation between democracies.' Thomas Regnier, European Commission spokesperson for tech sovereignty, echoed these sentiments, emphasizing that security concerns are a 'shared challenge' that transcends national borders and that solutions should not be 'discriminatory against partners.'

During a confidential G7 meeting, leaders from Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States explored the possibility of a 'trusted partner' framework for accessing advanced AI models, though specific details remain scarce. A similar tiered model for semiconductor chips, crucial for AI, was briefly implemented by the Biden administration in early 2025. This 'small yard, high fence' approach aimed to safeguard advanced US technology from nations like China and Russia but also generated 'disquiet' among allies, according to David Smith, an expert in US politics and foreign policy at the University of Sydney's United States Studies Centre.

While that scheme did not restrict allied access to semiconductors, it imposed limitations on their usage, including in commercial contexts or trade dealings with China. Smith noted the controversy, stating, 'Restricting access to advanced chips was very different,' as previous export restrictions typically targeted technologies with direct military applications. The Trump administration later rescinded this scheme in May 2025, subsequently approving the sale of the powerful Nvidia H200 chip to a select number of Chinese firms.

Europe Explores Independent AI Pathways

The Anthropic ban is now accelerating calls for greater self-reliance among US allies. Canadian Prime Minister Mark Carney remarked ahead of the G7 summit that the current situation with Mythos and Fable highlights the risks of over-reliance. He urged nations to learn from the incident and diversify their technological foundations. Bruno Retailleau, a former French minister and 2027 presidential candidate, asserted that Anthropic should serve as a 'wake-up call' for Europe, emphasizing that 'a nation that depends on others for its technology is a nation that can be unplugged overnight.'

Retailleau advocated for treating AI with the same strategic importance as nuclear power, viewing it as integral to national sovereignty. Tom Tugendhat, a former security minister, commented that the event was the 'inevitable result of technology shaping warfare so that sovereignty is more about code than cannons,' adding that continued dependence is incompatible with sovereignty. Marcin Jerzewski, head of the Taiwan office of the European Values Center for Security Policy, suggested that European companies might benefit from the Anthropic controversy.

Reports in April indicated that the German armed forces would not award contracts to Palantir, a US big data analytics company, due to concerns about granting private industry access to national data systems. Both German and French domestic intelligence agencies have also partnered with European companies over Palantir to avoid excessive reliance on US technology. Jerzewski highlighted that the Anthropic discussion has also brought attention to Paris-based AI startup Mistral, which he described as 'the EU’s only major homegrown frontier-model competitor.' He concluded that 'European governments are growing uneasy about their overreliance on US-controlled technologies,' suggesting that European companies could gain an advantage from the Anthropic incident.

Source: US export ban on Anthropic’s AI models further strains alliances