The UK government is assessing the potential economic and national security impact of losing access to frontier AI models, according to people familiar with the matter. The Cabinet Office has held discussions with the AI Safety Institute and the Department for Science, Innovation and Technology about what would happen if the UK were cut off from the most advanced AI systems.
The work is part of a broader effort to understand the risks and opportunities posed by frontier AI, which some experts fear could be used to create bioweapons or carry out cyber attacks. Whitehall officials have met with AI companies including OpenAI and Google DeepMind about the scenarios, the people said.
Dependence on foreign AI
Most of the world's leading AI labs are based in the US, including OpenAI, Google DeepMind and Anthropic, while China has emerged as a rival hub with companies such as Baidu and Alibaba. The UK has no home-grown frontier AI labs of its own, leaving it dependent on models developed elsewhere.
That dependence has raised questions about what would happen if access were cut off, either through geopolitical tensions, export controls or a decision by a company to withhold its technology. The assessment is examining both the economic costs and the national security implications of such a scenario, one person familiar with the matter said. It is at an early stage and has not yet produced formal findings.
The exercise is separate from work being done by the AI Safety Institute to evaluate the safety of specific models before release, the people said. Instead, it is a broader geopolitical exercise looking at the UK's reliance on AI developed in other countries, particularly the US. It also feeds into the government's work on AI for Government priorities, including the forthcoming national AI strategy due later this year.
Positioning as an AI safety hub
The UK has been trying to position itself as a hub for AI safety, hosting the world's first global AI safety summit at Bletchley Park in November. Prime Minister Rishi Sunak has made AI safety a priority of his government, and has said he wants the UK to be the "geographical home of AI safety".
The government has launched a £100mn AI taskforce to develop the UK's sovereign capabilities in the technology, including the creation of a British "foundation model" to rival those being built in the US. Officials have previously discussed whether the UK should build its own frontier AI model as a matter of national security, the people said. Such a project would be extremely expensive and would require access to vast amounts of computing power.
The economic assessment is being led by the Cabinet Office's resilience and emerging technology teams, working with the AI Safety Institute and the Department for Science, Innovation and Technology. It is expected to feed into the government's forthcoming AI strategy, which is due to be published later this year. The Cabinet Office declined to comment. The Department for Science, Innovation and Technology and the AI Safety Institute did not respond to requests for comment.
The work comes as governments around the world scramble to respond to the rapid advance of AI, which has been turbocharged by the release of OpenAI's ChatGPT in late 2022. Ministers have warned that frontier AI models - the most advanced systems, typically trained on vast amounts of data - could pose existential risks to humanity if not properly controlled. But the technology also offers large economic opportunities, and the UK has been keen to attract AI companies to set up in the country. Those working in government roles can track how these developments affect public-sector adoption through resources on Generative AI and LLM developments.
Why this matters for government professionals
This assessment signals that frontier AI access is now a strategic security issue, not just a technology policy matter. For civil servants and public sector officials, the practical question is how to plan for a scenario where the UK cannot rely on US-built models - which affects procurement decisions, data handling rules and contingency planning for critical public services that may come to depend on these systems. The outcome of this review could shape which AI tools government departments are allowed to use, and under what conditions.
Your membership also unlocks:
