Correspondence

AI sovereignty: preserving UK agency in an AI-enabled world (web version)

Published 19 October 2026

100 Parliament Street
London
SW1A 2BQ

The Rt Hon Kanishka Narayan MP
Cabinet Office
70 Whitehall
London
SW1A 2AS

‘Sent by email only’

30 July 2026

Dear Minister Narayan,

AI sovereignty: preserving UK agency in an AI-enabled world

On behalf of the Council for Science and Technology (CST), we would like to congratulate you on your appointment, and to offer our support in delivering your priorities.   

The Council is government’s top-level independent advisory body on cross-cutting science and technology issues. The Council brings together leaders across academia, business and finance to advise the Prime Minister and the Cabinet on how science and technology can support delivery of the government’s ambitions.  

We are developing advice for the government on artificial intelligence (AI) sovereignty. AI sovereignty is important because the choices the UK makes now will shape long-term competitiveness, resilience, public trust and strategic autonomy. The Council believes that government needs a clear view of where national capability matters, where dependence is acceptable, and where the UK has a credible opportunity to lead. 

We define AI sovereignty as having sufficient agency across the AI value chain to deliver the UK’s chosen outcomes, capture economic and public value for the UK, and maintain resilience under duress. Our advice recommends that government establishes clear strategy and governance arrangements to guide and coordinate choices on AI sovereignty, while investing in domestic capability where the UK has a credible advantage. Annex A provides a summary of the advice as it currently stands. 

We are grateful to Council members for developing this advice, in particular, Professor Sir Nigel Shadbolt, Avid Larizadeh Duggan, Mark Enzer, Priya Lakhani, Sir John Lazar, Professor Nick McKeown, Professor Andrew Morris, Professor Sarah Sharples, and Richard Slater.  

This letter is copied to the Prime Minister, the Secretary of State for Business, Innovation, Science and Trade, the Secretary of State for Defence, the Secretary of State for Digital, Culture, Media and Sport, the Chancellor of the Exchequer, the Chancellor of the Duchy of Lancaster, the Chair of the Prime Minister’s AI Taskforce, the Cabinet Secretary and relevant departmental permanent secretaries.  

Yours sincerely,

Dame Angela McLean

Co-chair

Lord Browne of Madingley

Co-chair

ANNEX A – EXECUTIVE SUMMARY OF CST’S ADVICE ON AI SOVEREIGNTY  

CST defines AI sovereignty as having sufficient agency across the AI value chain to deliver the UK’s chosen outcomes, capture economic and public value for the UK, and maintain resilience under duress. It depends on developing the right skills, organisations, data, practical experience and technical strengths. 

AI sovereignty is a whole-system challenge. It cannot be achieved by investing in technologies or projects in isolation. Without deliberate and urgent action, the UK risks becoming a high-value AI customer, rather than a country able to shape the technology and its usage. 

Skills, commercial value and capability will accumulate where AI systems are developed and used. The benefits from early action are substantial. Choices made now will determine whether we can build the future capability needed for prosperity and resilience. We are well placed to benefit from our strong research base and trusted institutions but need to decide which technologies to develop ourselves and which to securely access from other countries. 

The way for us to achieve AI sovereignty is to build trusted systems that deliver real public, economic and security benefits. That way we can keep more of the value in the UK and preserve future strategic choice. We make 4 recommendations to support this through joined-up investment, governance and delivery decisions. 

Recommendation 1: Develop a UK AI Sovereignty Plan based on a clear overview of the AI ecosystem. 

The first step of the plan should be rapidly mapping out the main parts of the AI ecosystem, including computing power, data, skills, models, governance and public adoption. The map should identify existing activity, dependencies and gaps. It should be produced quickly and used as a practical decision tool to identify which capabilities the UK should own, collaborate with others on, or secure resilient access to. 

These choices should inform a set of rapid AI pilots to be applied to public interest use cases such as national security, health, or education. These should include engagement plans that give the public a role in shaping their design.

Each pilot should answer 6 questions:

• what problems are to be solved?
• what sovereign data do we hold?
• what model capability is needed?
• what system must be built around it?
• what value can be captured?
• how will success be measured?  

Recommendation 2: Set up clear governance and coordination for AI sovereignty choices. 

Clear governance is needed to turn the AI Sovereignty Plan into action. Government should designate a single accountable owner, such as the head of the Cabinet Office AI Taskforce. This owner needs to bring departments together, and to shape funding and delivery choices. Governance should combine strategic and technical advice to shape the direction of the pilots and assess model, data and computing options. 

Recommendation 3: Invest now in the foundations of long-term UK AI capability.  

The foundations of AI sovereignty are not just models, but people, skills, organisations, data, trust and adoption. The UK needs attractive career pathways and opportunities for practical experience if expertise is to stay and grow in the UK. Government should offer an AI skills and secondment route that complements TechFirst. It should also strengthen the evidence base on AI adoption and productivity and continue supporting fundamental computing research including low-power approaches. 

Data is one of the UK’s most important assets, but is only valuable when it is managed, secured and used effectively. Government should identify important UK datasets, and test access arrangements in at least 2 high-value areas, such as health and creative data. 

Recommendation 4: Maintain trusted access to leading AI models while building selected UK capability.  

While access to the most powerful global models is important, the UK should not depend on them for all use cases. In some cases, a more specialised system developed domestically will work better. The UK must maintain access to leading global AI capabilities while building selected capabilities that strengthen independence, resilience and future choice.  

The AI pilots should be used to test how different types of AI systems (such as open weight models and domestic smaller models) can be used to address different requirements and tasks. Trialling different ways to build, train and access models will build enduring national skills, institutional knowledge and technical judgement. Government should allocate some national AI computing capacity for these trials and should consider proposals such as the Alan Turing Institute’s (ATI)^{[footnote 1]} for a public-interest sovereign base model.  

CST recommends these steps so the UK can benefit from AI and maintain our ability to make choices about how it is developed and used. CST can provide further independent support and advice as required.

1. The ATI have developed a proposal to deliver an open-source sovereign base model from scratch, which they would release as a public good. Their proposal presents the sovereign base model as national infrastructure: an open, UK-governed asset that could be used by researchers, start-ups, small- and medium-sized enterprises, public services and the open-source community, rather than a proprietary product for a single firm or department. ↩