The United States has imposed new restrictions on artificial intelligence models developed by Anthropic, a leading US-based AI company, prompting European Union officials to renew calls for greater technological sovereignty. The move, which limits the export and use of certain Anthropic models outside the US, has been framed by Washington as a national security measure. However, critics argue it is part of a broader effort to maintain American dominance in the rapidly evolving AI sector.
EU digital chief Margrethe Vestager said the restrictions 'prove that Europe cannot rely on third countries for critical AI infrastructure'. Brussels is now accelerating plans under the European Chips Act and the EU AI Act to foster homegrown AI development and reduce dependency on US technology. The EU has already invested billions into domestic AI research and is pushing for interoperable standards that would allow European firms to compete without being locked into US ecosystems.
For UK businesses, the development presents a dual challenge. The UK is not part of the EU AI Act but has signalled it will align closely with Brussels on AI safety and governance. At the same time, many British companies rely on US AI platforms for everything from customer service chatbots to data analytics. Industry experts warn that a fragmented regulatory landscape could increase compliance costs and limit access to cutting-edge tools.
'The UK is caught between two regulatory blocs,' said Dr Eleanor Shaw, a technology policy researcher at the University of Cambridge. 'If the US tightens export controls and the EU pushes for sovereignty, British firms could face higher costs and reduced choice. The government must decide whether to double down on the US alliance or forge a closer relationship with Brussels.'
The UK's Information Commissioner's Office (ICO) has yet to issue formal guidance on the implications for AI procurement. However, the regulator has previously warned that businesses must conduct due diligence on the data flows and security risks associated with foreign AI models. The ICO is expected to publish updated guidance on cross-border AI use later this year.
Economically, the restrictions could slow the adoption of AI among UK small and medium-sized enterprises, which often lack the resources to navigate complex export controls or develop proprietary models. The UK government has pledged £900 million to boost domestic AI supercomputing capacity, but experts say that without a clear regulatory path, the investment may not translate into commercial advantage.