Anthropic Model Shutdown Triggers EU Sovereignty Debate
After the US barred non-US citizens from Anthropic's Fable 5 and Mythos 5 models, European officials and researchers are divided on whether Europe should build its own AI or negotiate better access terms.
The European Commission is assessing the fallout from a US government export control order that forced Anthropic to pull its Fable 5 and Mythos 5 models for all non-US citizens worldwide, reigniting a debate over Europe’s dependence on American artificial intelligence infrastructure. [1]
Thomas Regnier, the Commission’s spokesperson for technological sovereignty, told Euronews that any emergency measures must “not be discriminatory against partners,” describing the situation as “a further illustration of why Europe needs to strengthen its technological sovereignty.” [1] He framed it as “a shared challenge, not one confined to a single jurisdiction or company.” [1]
Anthropic pulled the models following a US government order tied to national security concerns, and talks to restore access are ongoing. [1] Anthropic CEO Dario Amodei is scheduled to join other leading AI company heads at a working dinner with G7 leaders. [1]
European researchers, in statements published by the Science Media Center, broadly agreed the episode was a wake-up call but diverged sharply on the appropriate response. [1]
Thorsten Holz of the Max Planck Institute for Security and Privacy called it striking that a single foreign government order could “shut down a model overnight for all non-US citizens,” including European businesses. [1] He argued that digital sovereignty does not require self-sufficiency, but rather the ability to use critical technology even during geopolitical conflicts. [1]
Konrad Rieck of TU Berlin went further, warning that US models can be “shut off at any time, sometimes for opaque reasons,” and arguing that Europe must develop and operate its own capable models. [1] Gitta Kutyniok of LMU Munich called for an “Airbus moment” for AI, with coordinated investment in foundation models, chip design, and energy-efficient computing. [1]
Paul Röttger of the Oxford Internet Institute pushed back on that view, arguing that “Europe won’t be able to develop models like Mythos or Fable 5 in competition with the US.” [1] He said access should instead be secured through contracts tied to data center investments and backed by credible trade policy threats. [1]
Other researchers highlighted the structural obstacles to building homegrown capacity. Matthias Hein of the University of Tübingen warned that Europe would need not just one but several domestic providers, since commercial companies cannot be relied upon to keep releasing open-weight models. [1] Jonas Geiping of the ELLIS Institute Tübingen noted that French AI company Mistral has “fallen far behind” over the past two years, and that the basic infrastructure — large-scale data centers and sufficient power generation — is lacking, with Germany’s power generation having dropped back to 1985 levels. [1]
Geiping also cautioned against treating the situation as analogous to nuclear weapons tensions, a comparison Anthropic itself has drawn. [1] Unlike nuclear weapons, he argued, AI is deeply embedded in the economy, meaning a shutdown during a diplomatic conflict could cause serious economic damage if business processes cannot function without access to capable AI systems. [1]
Sources
This article was drafted with AI from the cited sources and checked against them before publication. Spot an error? Let us know.
Apple says DMA interoperability rules make it impossible to launch its new AI-powered Siri in Europe; the European Commission says nothing in the law prevents Apple from doing so.
Apple has announced that its new AI-powered Siri will not launch on iPhones and iPads in the European Union, citing the bloc’s Digital Markets Act as the reason — a claim the European Commission flatly disputes. [1]
The DMA requires platforms designated as gatekeepers to give competitors the same kinds of data access they themselves enjoy, with limited exceptions. [1] For an AI assistant designed to access apps, personal information, photos, messages, and videos and take actions on users’ behalf, Apple argues that means handing over an enormous amount of system access to outside companies including OpenAI, Google, and Anthropic. [1]
Apple says such access would compromise the privacy and security of its customers. [1] The company proposed an intermediary mechanism it calls the Trusted System Agent, which it said would give rival AI agents comparable access and capability without direct system exposure, and said it would need 18 months to implement it on a “gradually rolling” basis. [1] Apple says the European Commission rejected that proposal and its others, leaving the company to declare that “there is currently no timeline for Siri AI’s availability in the EU on iOS and iPadOS.” [1]
Apple’s senior vice president of marketing Greg Joswiak, speaking at a press briefing, said the EU’s interpretation of the DMA would require giving others “unfettered access” to practically everything on a user’s device. [1] He added that the Commission has not “meaningfully engaged with us in our proposals.” [1]
The European Commission offered a direct rebuttal. “Nothing in the DMA prohibits Apple from introducing new products and services in the EU,” Commission spokesperson Ricardo Cardoso told The Verge. [1] Cardoso also stated that “Apple did not develop proposals for DMA compliant interoperability solutions,” a characterization Joswiak disputed. [1]
Apple took the unusual step of dedicating part of its WWDC 2026 keynote to explaining the European situation and published a blog post titled “Due to DMA, Siri AI delayed in EU for iOS 27 and iPadOS 27.” [1] China will also not receive Siri AI due to regulatory challenges, a fact Apple disclosed in a one-sentence footnote. [1]
Experts contacted by The Verge offered mixed assessments. Friso Bostoen, a professor of competition law and digital regulation at Tilburg University, acknowledged real security and privacy risks in forcing platforms to open their systems, but said Apple’s privacy arguments have not always withstood judicial scrutiny, pointing to recent court cases in the UK and US. [1] Jan Penfrat, a senior policy adviser for European Digital Rights, characterized Apple’s moves as “a lobbying tactic” aimed at pressuring the Commission to allow it to sidestep the DMA. [1]
Michael Veale, a professor of technology law and policy at University College London, argued that Apple is already making an exception to its own privacy model by giving Siri AI broad access across apps, but resists extending equivalent access to competitors. [1] Veale and Penfrat both noted that Apple has not made its proposed Trusted System Agent solution public, making independent assessment impossible. [1] Bostoen separately questioned why Apple would need 18 months to implement interoperability measures that were foreseeable and could have been developed alongside Siri AI itself. [1]
The dispute is not the first time Apple has cited DMA interoperability requirements to withhold features from European users; the company has previously attributed the absence of AirPods live translation, iPhone mirroring, and certain Maps features in the EU to the same law. [1]
Sources
This article was drafted with AI from the cited sources and checked against them before publication. Spot an error? Let us know.