The Trump administration is reportedly preparing to reject the United Kingdom’s request for an exemption from U.S. restrictions on Anthropic’s most advanced AI models, Fable 5 and Mythos 5. The controls were introduced after U.S. officials raised national security concerns that the models could potentially be “jailbroken” to expose advanced cybersecurity capabilities. British officials, including Prime Minister Keir Starmer, have lobbied Washington for a carve-out, but sources indicate there is “zero chance” of a UK-specific exemption.
The dispute has sparked wider debate about what some observers call an AI “kill switch” risk—the possibility that countries, companies, or critical sectors relying on foreign AI providers could suddenly lose access to essential technology. Because Anthropic was required to block access for non-U.S. users, organizations across Europe were reminded that key AI infrastructure remains controlled by a small number of American firms and can be restricted through government action.
Anthropic has challenged the government’s assessment, arguing that the cited vulnerability was narrow and did not justify withdrawing frontier AI models from international users. The company warned that applying similar standards across the industry could significantly slow or halt the release of future advanced AI systems. Emergency discussions between Anthropic and U.S. officials have reportedly taken place, but no resolution has yet been announced.
Beyond the immediate dispute, the episode has become a wake-up call for Europe and other U.S. allies. Policymakers are increasingly emphasizing “sovereign AI” capabilities, arguing that nations need greater control over their own AI infrastructure, computing resources, and foundation models to avoid dependence on foreign providers. The controversy has intensified calls for investment in domestic AI ecosystems and international partnerships that reduce reliance on a handful of U.S.-based companies.
