French presidential hopeful Gabriel Attal is making as many headlines for his supposed calls to “reinvent Macronism” (economic principles advocated by Emmanuel Macron) as he is for his assertion that “the AI war has already begun” right now.
Attal’s comments come in the wake of the U.S. government’s order to bar non-U.S. citizens from accessing Anthropic’s most powerful AI model for cybersecurity research.
Transatlantic Tribulations?
Although this mandate appears to be slightly woolly in some respects – many U.S. organizations and software engineering communities will naturally straddle transatlantic connections, so the ruling could be tough to enforce – the development itself is fuelling calls throughout Europe to rely less on U.S. technology and push for greater European sovereignty.
Anthropic’s original statement confirmed that the U.S. government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
Fable 5 and Mythos 5: Abruptly Disabled
“The net effect of this order is that we must ‘abruptly disable’ Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Anthropic models will not be affected,” stated Anthropic.
The company says that its “understanding” is that the government believes it has become aware of a method of bypassing, or “jailbreaking,” Fable 5.
“We reviewed a demonstration of this specific technique being used to identify a small number of previously known, minor vulnerabilities. These vulnerabilities all appear relatively simple, and we have found that other publicly available models are able to discover them as well without requiring a bypass,” added Anthropic.
Back Over The Pond
Back across the pond in Europe, the EU executive has said that existing (and currently developed) EU laws covering AI and cybersecurity have “served as a good basis” in terms of their ability to govern Europe’s protection against cyber risks and threats emanating from AI-powered actors.
A European Commission spokesperson explained the EU’s central position here, saying that, “We continue discussions with our allies on these models. It is in our mutual interest to tackle these risks and strengthen cybersecurity.”
Keen to wrestle itself away from U.S. AI dependency (and datacenter-related, microprocessor chip-related, enterprise application-related and wider IT stack-related dependencies at a number of levels right now), the EU has been mustering support for its calls to sever (or at least partially release) ties with the U.S. at this level.
“The issue here is not whether the U.S. is right to take national security concerns seriously. Of course it is. The issue is that Europe cannot keep building its tech stack on access that can be switched off overnight by a foreign government,” Wrote Finnish MEP Aura Salla on LinkedIn.
Salla said that she recognizes that the “key security concerns” around Mythos 5 and Fable 5 are not “normal” software vulnerabilities but are concerns more closely aligned with the likelihood of bypassing their safeguards or being able to obtain responses that would normally be restricted.
“Another controversy concerns that the model can secretly provide weaker assistance for AI development work to reduce the risk of creating other powerful systems. Lack of transparency here is a trust concern,” she added.
Oh, Canada!
In Canada, ex-Bank of England director and now Canadian Prime Minister Mark Carney has said that American restrictions on Anthropic’s newest AI models “show the dangers of overreliance” on a restricted or limited group of American providers when it comes to the AI model market.
“The situation we’re in collectively right now with Mythos and Fable is something that can happen with overreliance on certain models. Nobody has done anything wrong in the situation. But we will have done something wrong if we just accept this, don’t take the lesson, don’t build out and diversify,” stated PM Carney. “You’ll hear me say this over and over again. It is never a good idea to have one option.”
Eugene Cheah, CEO and co-founder of Featherless.ai, has commented that trying to control AI by restricting access to closed models could backfire for Anthropic. It may push more governments, companies and developers towards open models they can run, adapt and keep access to themselves.
“That is the point of open source – research is not limited to a small number of labs, and teams can tune models for the problems they are trying to solve. Rio 3.5 Open, an open model released by the City of Rio de Janeiro, is a good example. It shows how far open model work can be pushed when it is adapted for a specific use case,” said Cheah. “Closed models will still be ahead in some areas, especially on the hardest and most expensive tasks. But most organisations do not need a Porsche for every journey. They need models that are good enough for the job, that they can keep access to and that are not dependent on decisions made by a handful of companies.”
Cheah says that if AI is going to affect every part of business and public life, it can’t be controlled by a small number of providers – he thinks that the rest of the market will simply not accept this.”
What To Think Next
The restriction of Anthropic’s Fable 5 and Mythos 5 models may be taken by many as a stark wake-up call for global tech independence, exposing the volatility of relying on a foreign-controlled IT infrastructure.
While the United States views the restriction as a necessary national security measure to prevent AI safeguard bypasses, the move has inadvertently accelerated a global push for technological sovereignty. By demonstrating that access to critical AI tools can be revoked overnight, this decision leaves Europe and international allies with a clear ultimatum: Either accept the strategic vulnerability of overreliance on American providers, or aggressively invest in diversifying and building out their own domestic tech stacks to secure their digital futures.
