Under strict mandatory directives from President Donald Trump's administration, Washington has imposed unprecedented restrictions that prevent non-U.S. citizens — including European allies and foreign corporate employees — from accessing the latest leading AI models.

The protectionist move raises deep fears in Brussels and Paris that foreign technology could become a tool of blackmail and strategic influence, amid a technical divide between the imperative of national security control and the risk of stifling global innovation that enhances cyber defenses.

Emerging U.S. Restrictions on European Allies

European academics and politicians have criticized the new U.S. restrictions on advanced AI models, arguing that they represent a dangerous shift in the relationship between citizenship and technology, and a redefinition of who is entitled to participate in the advanced digital economy.

The scope of protectionist restrictions imposed by Washington was not limited to countries classified as strategic adversaries of the United States such as China or Russia, but also abruptly extended to citizens of European countries and Washington's traditional allies.

According to European media, these angry attitudes came in the context of the US administration's decision to restrict non-US access to Antropic's latest models, known internally as "Mythos 5" and "Fable 5", where they were disabled and access restricted to prevent their use by non-US citizens, including the company's own foreign employees, in compliance with a direct directive issued by the Trump administration.

Prominent European legal platforms have published analyses that have described the decision as a "future ban", noting that European citizenship – if not coupled with US citizenship or a special exception – has become an obstacle to access technologies that are at the heart of global planning and technological development.

France warns of foreign technology as security weakness

In political reactions, former French Prime Minister and presidential candidate Gabriel Attal warned that the world is entering an escalating "technological war" that reflects the intensifying geopolitical competition over control of the most readable technologies of the future.

Attal said that what the world is witnessing today "cannot be considered just a normal technical organization," stressing categorically that "the AI war has already begun."

The new U.S. restrictions are concrete evidence that Europe's continued reliance on foreign technology could at any moment turn into a fatal political and security weakness for the old continent, voices that have described the incident as "the beginning of a new technological cold war" that is reshaping the rules of competition in a sector that will determine the balance of economic and security power in the coming decades.

Washington's justifications

On the other hand, Washington justified its mandatory decision with categorical national security considerations, citing intelligence and technical concerns that the "Mythos 5" and "Fable 5" models possess highly sophisticated capabilities in detecting cyber vulnerabilities and developing sophisticated offensive tools that may be used to launch large-scale cyberattacks on critical infrastructure.

This is the broadest and most drastic step to date in restricting access to leading systems.

For its part, Anthropic expressed its official objection to the mechanism of implementing the decision, stressing that it received the government guidance without sufficient clarification or explanation of the security foundations on which it was based, describing the measure as not in line with the principles of transparency and technical clarity in place.

The company said it could be the result of a regulatory "misunderstanding", and said it hoped to restore access to the models as soon as possible for its crews and European counterparts.

The administration has defended its approach through a new regulatory framework aimed at vetting advanced "leading models" (such as ChatGPT , Claude, and Llama that are capable of inference and use tools independently) before launch, with the possibility of sharing them with the government in advance for risk assessment, although this framework, under its current texts, remains voluntary for companies and does not impose a comprehensive mandatory licensing system.

Warnings of the Chinese Rise

In the wake of the crisis, more than 100 cybersecurity experts — including professionals working for two tech giants Nvidia and Adobe — have called on the U.S. administration to immediately ease the restrictions.

In a joint statement, the experts warned that denying Washington's allies and specialists access to these models could weaken Western collective defense capabilities more than protect them, stressing that rival countries, led by China, are only a few months away from developing similar technological capabilities, and may already have access to more sophisticated systems than internally advertised.

Academic experts point out that this conflict exemplifies what is known as the "evidence dilemma": the slow action of governments leads to existential security risks, while excessive militancy and nationalist prohibition stifle innovation. These developments come in light of a growing gap in the global governance of AI and the absence of a binding international framework, despite several international AI safety summits being held in various capitals in recent years.