Artificial intelligence has become the latest issue to drive a wedge between the United States and its allies after President Donald Trump ordered tech giant Anthropic to cut off foreign access to its powerful Mythos-5 and Claude Fabre-5 AI models, citing national security concerns.
Last week, the United States issued an unprecedented order to all foreign nationals in and outside the United States, prompting Anthropic to take its two AI models completely offline to ensure compliance.
Recommended stories
list of 4 itemsend of list
Anthropic has granted 200 institutions in 15 countries access to Claude Mythos Preview, a frontier model for testing vulnerabilities.
Two public versions of this model, Mythos 5 and Fable 5, were scheduled to be released in early June.
Anthropic said the U.S. government did not give a reason for the order, but said it “understands” that the Trump administration believes it has figured out a way to “jailbreak” Fable 5.
The Trump administration’s ban immediately sent shock waves across Europe, which relies heavily on U.S.-developed AI.
French President Emmanuel Macron told the Group of Seven (G7) this week that the Trump administration’s order was a “wake-up call” about the dangers of AI, but that restrictions were “a bad thing”.
“The reaction is in a sense completely nationalist,” Macron said on Wednesday.
While the United States has targeted adversaries such as China and Russia with numerous technology regulations, the Trump administration’s Humanity Order applies equally to allies with which the United States has intelligence-sharing and mutual defense agreements.
The decision was a first for the AI industry, but followed other trade policy moves by the Trump administration.
Over the past 18 months, President Trump has launched a global trade war, annexed the autonomous Danish territory of Greenland and threatened to withdraw from the 77-year-old North Atlantic Treaty (NATO) alliance.
He also threatened to cut off arms supplies to Ukraine unless European allies cooperated in reopening the Strait of Hormuz. The Strait of Hormuz was effectively blocked by Iran after the US and Israel launched a war against Iran on February 28th.
America’s allies are now waking up to the realization that they are “currently far too vulnerable to the American tech-industrial complex,” Dex Hunter Tricke, director of the Center for Tomorrow, told Al Jazeera.
President Macron also emphasized the need for countries to work together to address AI issues and warned of the dangers posed by “non-cooperation between democracies.”
Macron’s comments were echoed by Thomas Renier, the European Commission’s technology sovereignty spokesperson, who told Al Jazeera that addressing security concerns is a “common challenge, not limited to a single jurisdiction or country.”
The solution also must not be “discrimination against partners,” Renier said.
In a closed-door meeting, the leaders of the G7 countries – Canada, France, Germany, Italy, Japan, the UK and the US – discussed potential “trusted partner” plans for access to cutting-edge AI models, but few details were revealed.
The United States briefly introduced a similar tiered model for semiconductor chips powering AI in early 2025 during President Joe Biden’s final weeks in office.
The plan, known in public policy circles as the “small garden, high fence” model, was aimed at keeping cutting-edge American technology away from countries like China and Russia, but it also created a degree of “anxiety” among allies, said David Smith, an expert on U.S. politics and foreign policy at the Center for American Studies at the University of Sydney.
The plan does not aim to restrict allies’ access to semiconductors, but rather to limit what they can do with those chips, including in commercial settings and trade deals with China.
“This was controversial because in the past, when the United States has restricted the export of certain technologies, they have always had direct military use,” Smith told Al Jazeera.
“Restricting access to advanced chips was completely different.”
The plan was scrapped by the Trump administration in May 2025, which subsequently allowed the sale of powerful Nvidia H200 chips to a limited number of Chinese companies.
Meanwhile, the ban on human settlements has accelerated calls for greater independence among U.S. allies.
Canadian Prime Minister Mark Carney told the media ahead of the G7 summit that “the situation we’re collectively in right now in Mythos and Fables is what can happen through over-reliance.”
“No one has done anything wrong in this situation. But if we just accept this and don’t learn the lessons and don’t build and diversify, we’ll have done something wrong.”
Bruno Letailault, a former French minister and 2027 presidential candidate, said Antropic should be a “wake-up call” for Europe: “A country that depends on other countries for technology is a country that can cut off electricity overnight.”
“We have to treat AI the same way we treated nuclear power. We have to consider AI as part of our sovereignty. Master it or live with it. There is no other way,” Lutailot told Saturday’s X, the day after the ban came into force.
Tom Tugendhat, a former security secretary under Conservative Chancellor Liz Truss and Chancellor Rishi Sunak, said the incident was an “inevitable consequence of technology shaping wars, as sovereignty is more about codes than it is about artillery”.
“We cannot continue like this and maintain our sovereignty,” Tugendhat said on Saturday’s X show.
Marcin Dzierszewski, director of the Taiwan office at the Center for Security Policy and European Values, said European companies could have an advantage in the humanity debate.
Prior to the artificial ban, German media reported in April that the country’s military would not sign contracts with Palantir, Peter Thiel’s big data analytics company that partners with law enforcement, defense and intelligence agencies, citing concerns about giving private companies access to national data systems.
Domestic spy agencies in Germany and France have also recently partnered with European companies over Palantir to avoid becoming overly dependent on U.S. technology, officials said.
As Anthropic fuels the debate about the need for AI sovereignty, it is also drawing attention to Paris-based AI startup Mistral, “a major competitor to the EU’s only homegrown frontier model,” Dzierszewski told Al Jazeera.
“Overall, European governments are becoming increasingly concerned about their over-reliance on U.S.-controlled technology,” he said.
“To this end, European companies may benefit from the Anthropic case.”
