The recent order by the Trump administration for the tech company Anthropic to restrict access to its advanced AI models, Mythos 5 and Claude Fable 5, has ignited tensions between the United States and its allies. The directive, issued last week, aims to address national security concerns and prohibits foreign nationals from accessing the AI systems both within and outside the US. As a consequence, Anthropic has taken the decision to take the two models offline to ensure compliance.
Previously, Anthropic had allowed 200 institutions in 15 different countries to engage with the Claude Mythos Preview model for vulnerability testing. The anticipation for the public release of Mythos 5 and Claude Fable 5, which was set for early June, has now been overshadowed by this development. Anthropic has expressed that the US government did not specify the reasons behind this sweeping ban, yet it was suggested that officials may have identified a method for “jailbreaking” Fable 5.
This sudden shift has sent shockwaves across Europe, where reliance on US-developed artificial intelligence technologies is significant. French President Emmanuel Macron labeled the US’s decision as a “wake-up call” regarding the inherent dangers of AI and criticized the unilateral approach as being “strictly nationalist.” Notably, the ban applies not only to adversaries like China and Russia but also affects allied nations with established intelligence-sharing agreements and mutual defense pacts with the US.
Experts are now raising alarms about the vulnerability of US allies amidst what is perceived as an over-dependence on American technological prowess. Dex Hunter-Torricke, president of the Center for Tomorrow, commented on the current situation, emphasizing that allies must recognize how exposed they are to the US techno-industrial complex. Macron underlined the necessity for collaborative efforts in addressing AI-related issues, warning that non-cooperation among democracies could pose significant risks.
During a recent closed-door meeting of G7 leaders, discussions revolved around creating a “trusted partner” scheme for access to elite AI technologies. While specific details remain under wraps, this model aims to ensure that sensitive capabilities are selectively shared among allies. A similar tiered approach had been introduced for semiconductor technology in early 2025, designed to control advanced US tech from reaching countries like China and Russia. However, these policies have been met with unease from allies who feel constrained.
Canadian Prime Minister Mark Carney voiced concerns that this scenario underscored the pitfalls of over-reliance on technology controlled by a singular nation. He stressed that it is crucial for allies to learn from this episode, advocating for diversification and self-sufficiency in technological capabilities to avoid future vulnerabilities.
In a similar vein, Bruno Retailleau, a former French minister, articulated that the Anthropic situation serves as a critical reminder for Europe about the dangers of dependence on outside technology, urging EU nations to treat AI development as a matter of national sovereignty. The sentiment that technology, especially AI, should be approached with the caution traditionally reserved for nuclear power resonated throughout discussions.
This contextual fallout from the Anthropic ban has prompted European nations, particularly in intelligence and defense sectors, to reconsider their partnerships with US tech firms. Reports suggest that countries like Germany and France are increasingly opting for domestic solutions to avoid becoming overly reliant on American technologies.
The complexities arising from the Anthropic order and its implications for international tech collaboration highlight a growing sense of urgency among allies to rethink their strategic approaches to AI development. As stakeholders gauge the potential impact, the conversation about AI sovereignty and technological independence is set to intensify in the coming months.
