The UK government has been urged to develop a national AI sovereignty strategy to strengthen domestic artificial intelligence capabilities, protect strategic interests, and ensure the country captures greater economic value from AI innovation. The Council for Science and Technology (CST) warns that without coordinated action, the UK could become a major customer of AI technologies developed elsewhere rather than helping shape their future.

The recommendations were outlined in a letter addressed to Kanishka Narayan MP, who was appointed Minister for Artificial Intelligence in July. Signed by CST co-chairs Dame Angela McLean and Lord Browne of Madingley, the letter was also copied to the Prime Minister, Chancellor of the Exchequer, and other senior government officials.

As the UK expands its adoption of artificial intelligence across public services and the wider economy, the Council argues that decisions made today will influence the country’s long-term competitiveness, technological resilience, public trust, and strategic independence.

What Is AI Sovereignty and Why Does It Matter to the UK?

AI sovereignty refers to a country’s ability to maintain sufficient control and influence across the artificial intelligence value chain to achieve its national objectives. This includes access to computing infrastructure, data, AI models, skilled professionals, governance frameworks, and the systems required to deploy AI effectively.

According to the CST, AI sovereignty means having enough agency across the AI ecosystem to deliver the UK’s chosen outcomes, capture economic and public value, and remain resilient during periods of disruption.

The Council stresses that sovereignty does not necessarily require the UK to develop every AI technology independently. Instead, the government needs to identify which capabilities are strategically important to build domestically, which technologies can be accessed securely from international partners, and where external dependencies may create unacceptable risks.

Dame Angela McLean and Lord Browne of Madingley warn that the UK could otherwise become a high-value AI customer instead of a country capable of influencing how the technology is developed and deployed.

The letter also highlights the economic implications of this challenge. Skills, commercial opportunities, and technological expertise tend to accumulate in countries and organisations where AI systems are developed, owned, and widely used.

For the UK, strengthening domestic AI capability could therefore help retain more economic value while supporting innovation, public-sector efficiency, and long-term technological resilience.

Four Key Recommendations for a UK AI Sovereignty Strategy

The Council for Science and Technology sets out four main areas for government action to establish a more coordinated approach to AI sovereignty.

1. Develop a National AI Sovereignty Plan

The CST recommends creating a comprehensive UK AI Sovereignty Plan that maps the country’s existing AI ecosystem and identifies critical capability gaps.

The assessment should cover computing infrastructure, data resources, workforce skills, AI models, governance arrangements, and public adoption. It should also distinguish between areas where the UK has strong domestic capabilities and those where international dependencies could pose strategic challenges.

The Council further recommends launching rapid AI pilot projects in areas of public interest, including healthcare, education, and national security. Public engagement should form part of these initiatives to help ensure that AI deployment addresses real needs and maintains public confidence.

2. Establish Clear AI Governance and Accountability

Effective implementation will require clear leadership across government departments and public institutions.

The CST recommends appointing a single accountable owner to coordinate the national AI sovereignty agenda. One potential option identified in the letter is the head of the Cabinet Office AI Taskforce.

A central point of accountability could help align policy priorities, reduce fragmented decision-making, and ensure that AI investments contribute to shared national objectives.

Clear governance would also help the government assess risks, monitor progress, and coordinate initiatives across the wider public sector.

3. Invest in UK AI Skills, Research, and Data Infrastructure

Long-term investment in domestic AI capability is another central recommendation.

The Council proposes developing an AI skills and secondment pathway that complements the TechFirst initiative. It also calls for a stronger evidence base to understand how businesses and public institutions adopt AI and whether that adoption translates into measurable productivity improvements.

Research investment should include fundamental computing research, particularly low-power computing approaches that could support more efficient AI infrastructure.

Data management is another priority. The CST recommends identifying strategically important UK datasets and testing practical access arrangements in at least two high-value areas, such as healthcare and creative industries.

Improving access to valuable datasets, while maintaining appropriate security and governance, could help researchers, businesses, and public organisations develop AI applications that address UK-specific challenges.

4. Maintain Trusted Access to Leading AI Models

The Council recommends balancing domestic AI development with secure access to leading international AI models.

Rather than attempting to build every capability independently, the UK should develop selected areas of domestic expertise while avoiding excessive reliance on global models for every application.

This approach would allow the country to benefit from international innovation while retaining greater flexibility over strategically important systems and use cases.

The CST also points to the UK’s established research base and trusted institutions as potential advantages in developing a distinctive and resilient AI ecosystem.

Six Questions to Evaluate Every UK AI Pilot Project

The CST recommends applying a consistent set of questions to AI pilot projects to ensure that each initiative supports national priorities and delivers measurable value.

The six questions are:

1. What sovereign data does the UK hold? Identify relevant datasets and assess how they can be accessed and used securely.
2. What problem needs to be solved? Define the specific public-sector, economic, or strategic challenge the project will address.
3. What AI model capability is required? Determine the technical capabilities needed to achieve the intended outcome.
4. What wider system must be built? Consider the infrastructure, processes, governance, and workforce required to deploy the technology.
5. What value can the UK capture? Assess potential economic benefits, public-service improvements, and opportunities to strengthen domestic expertise.
6. How will success be measured? Establish clear performance indicators to evaluate results and guide future investment.

This framework is intended to encourage a practical, outcome-driven approach to AI sovereignty rather than focusing on technology development in isolation.

Why Data Management Is Central to AI Sovereignty

The CST emphasises that AI sovereignty is a whole-system challenge that cannot be achieved through isolated technology investments.

Although advanced AI models attract significant attention, the Council argues that a country’s long-term capabilities also depend on its people, skills, organisations, data resources, public trust, and ability to adopt new technologies effectively.

Data is particularly important because access to high-quality information can influence the development, performance, and usefulness of AI systems.

However, simply possessing valuable datasets is not enough. Data must be managed properly, protected against misuse, and made available through appropriate access arrangements if it is to support meaningful innovation.

For the UK, this means treating data governance and infrastructure as essential elements of national AI policy. Stronger data management could help public institutions and businesses develop more relevant AI applications while maintaining security, accountability, and public confidence.

The Council’s recommendations also suggest that successful AI deployment requires coordination between technical development and the organisations expected to use the resulting systems.

UK AI Strategy: Balancing Domestic Innovation and Global Partnerships

The proposed AI sovereignty plan does not call for complete technological independence. Instead, it highlights the need for a balanced strategy that combines domestic capability with trusted international access.

For technologies considered strategically important, the UK may need to strengthen its own research, infrastructure, workforce, and development capacity. In other areas, secure partnerships and access to international AI systems may offer a more practical route.

Determining where to invest, where to collaborate, and where to reduce dependency will be central to the government’s approach.

The CST believes the UK can build on its research strengths and established institutions, but it must translate those advantages into practical capabilities that support economic growth, public services, and national resilience.

What Happens Next for the UK’s AI Sovereignty Plan?

The CST’s letter sets out recommendations rather than a fully costed implementation programme. It does not specify funding commitments, delivery deadlines, or a detailed timetable for introducing the proposed measures.

The next steps will therefore depend on how the government responds to the recommendations, assigns responsibility, and determines which capabilities require priority investment.

Progress will also depend on the government’s ability to coordinate public-sector AI initiatives, improve access to strategically important data, develop specialist skills, and measure the real-world impact of AI adoption.

Ultimately, the proposed UK AI Sovereignty Plan seeks to ensure that the country does more than consume artificial intelligence developed elsewhere. By combining targeted investment, clear governance, secure access to leading models, and effective data management, the UK could strengthen its ability to influence AI development and capture more of its economic and public value.