The U.K AI Plan For Change

Kier Starmers AI Speech on 13th January 2025

Britain’s AI Opportunity

Today, the UK ranks as the world’s third-largest market for Artificial Intelligence, hosting an impressive network of international talent and pioneering AI companies such as Google, DeepMind, ARM, and Wayve. Our legacy of scientific discovery—from Alan Turing’s groundbreaking work on algorithms to Tim Berners-Lee’s invention of the World Wide Web—has forged a rich heritage in computing. Yet despite this strong foundation, Britain now faces stiff competition from the United States and China, whose rapid AI advances could leave the UK behind.

We believe the next phase of AI development is our chance to redefine our place in the global tech arena. The UK has a responsibility to lead in responsible AI development, just as we have done with AI safety. That’s why one of my earliest acts as Secretary of State for Science, Innovation and Technology was to commission Matt Clifford to design a comprehensive AI Opportunities Action Plan for the British government.

Seizing the Next AI Revolution

This Action Plan reveals how we can shape AI’s future under the principles of a modern social market economy. By partnering with leading AI firms, our world-class academic institutions, and aspiring AI entrepreneurs, we aim to direct AI’s transformative power toward collective prosperity, stronger public services, and expanded personal opportunities. Our vision ensures:

  1. AI-Driven Economic Growth: We rely on a thriving economy to support both our people’s prosperity and the quality of public services.
  2. Direct Benefits to Working People: From healthcare to education, AI should elevate how citizens interact with government and public services.
  3. New Opportunities at Work: As AI becomes more common in the workplace, it must unlock novel career paths and possibilities, instead of merely disrupting traditional roles.

Government’s Commitment to AI

Our decisive actions already include:

  • Planning Reforms: Making it simpler to build the data centres powering the AI era.
  • Skills England: Preparing UK citizens for the roles of tomorrow in AI-fuelled industries.
  • Digital Centre of Government: Accelerating technological innovation across the public sector so citizens receive a smoother, more efficient experience.

The proposals in this plan are bold and unapologetic. Government must match that ambition. Realizing our AI vision demands effort, tough decisions, and a cooperative bond between the public and private sectors. This is our roadmap for taking on the challenge—starting now.

The Rt Hon Peter Kyle MP
Secretary of State for Science, Innovation and Technology

The Opportunity

AI capabilities are evolving at an astounding pace. Should this trend continue, AI may become the single most powerful tool for the government to achieve its five key missions, notably the push for inclusive economic growth. Meeting our ambition of the highest sustained growth in the G7 is nearly inconceivable without leveraging AI’s potential.

A successful AI strategy begins by examining our strengths and weaknesses. Fortunately, the UK’s foundation is both solid and world-leading in crucial areas:

  • Exceptional AI Research: A steady pipeline of top-tier talent emerges from our globally respected universities.
  • Dynamic Startup Scene: We have a flourishing ecosystem of skilled entrepreneurs and rising levels of sophisticated investment.
  • Major AI Players: London-based Google DeepMind, plus offices for OpenAI, Anthropic, Microsoft, and Meta, along with homegrown successes like Wayve.
  • Global Leadership in AI Safety and Governance: Through the AI Safety Institute and a flexible regulatory climate.

Still, if we want to remain in the forefront, we must lead in both developing and applying AI. Our goal is a thriving AI ecosystem populated by influential UK-based companies across the entire “AI stack” and widespread AI adoption throughout all sectors.

Pursuing that aim demands big thinking. The government must:

  1. Invest in AI Foundations – Establish world-class computing infrastructure, data resources, accessible talent, and agile regulations.
  2. Accelerate AI Adoption – Quickly pilot AI tools in the public sector and empower the private sector to do the same, improving services for citizens and boosting productivity.
  3. Be an AI Maker, Not a Taker – As AI becomes more potent, the UK should be the best partner for those forging cutting-edge technology and cultivate homegrown AI champions.

These three goals form the structure of this Action Plan. Although no one knows exactly what AI will be like in a decade, the danger of underinvesting is far more serious than the risks of overcommitment. If growth in AI capabilities holds steady, being at the forefront means the UK can help shape scientific and technological progress and reap the shared rewards.

1. Lay the Foundations to Enable AI

1.1 Building Ample, Secure, and Sustainable Infrastructure

Modern AI owes its astonishing growth to massive increases in computational power (compute). High-performance data centres are indispensable for training models and running AI applications at scale.

The UK needn’t own every single computer that powers its AI sector, but ensuring sufficient and secure compute is critical to our national interests. This includes:

  • Sovereign AI Compute – Publicly overseen resources providing independence for research and critical services.
  • Domestic Compute – Privately owned but UK-based infrastructure. This secures our economic future, fosters jobs, and encourages further AI-based services.
  • International Compute – Strengthening ties with partners who share our values and building reciprocal alliances for shared research and computational power.

Key Actions:

  1. Publish a 10-Year AI Infrastructure Plan within six months, supported by a dedicated investment framework.
  2. Expand AI Research Resource (AIRR) Capacity by 20x by 2030, starting immediately, to train multiple models yearly and retain world-class capability.
  3. Appoint Mission-Focused AIRR Directors to allocate compute effectively to high-value projects aligned with national priorities.
  4. Establish ‘AI Growth Zones’ (AIGZs) to fast-track data centre development, focusing on regions ripe for rejuvenation and with existing energy capacity.
  5. Minimize Environmental and Security Risks of compute infrastructure, while exploring advanced hardware solutions.
  6. Forge International Compute Partnerships with like-minded allies, building on initiatives like the EuroHPC Joint Undertaking.

1.2 Unlocking Data Assets in Public and Private Sectors

High-caliber data fuels the next wave of AI breakthroughs. The UK should unlock key public datasets—especially those that provide valuable insights not yet included in standard training sets. We must also encourage private-sector collaborations to unleash innovation and attract global investment.

Key Actions:

  1. Identify 5 High-Impact Public Datasets for release to researchers and innovators under clear guidelines.
  2. Shape the Future of Data Collection, going beyond mere availability to strategically gather datasets in areas of proven UK excellence (e.g., genomics, disease recognition).
  3. Publish Open Data Standards to guide government departments on releasing datasets suitable for AI, under robust privacy and security protocols.
  4. Pair Compute with Proprietary Data as part of an attractive package for startups and researchers to base themselves in the UK.
  5. Invest in Public Sector Data Infrastructure so that new, high-value data emerges readily for advanced AI training.
  6. Reward Creation of Private Datasets, especially those beneficial to the research community, potentially in partnership with UKRI.
  7. Develop a Copyright-Cleared British Media Dataset, by collaborating with national cultural institutions, to license internationally for AI training.

1.3 Training, Attracting, and Retaining World-Class AI Talent

To build a vibrant AI ecosystem, the UK must educate and attract tens of thousands of AI professionals and proactively increase its global share of top-tier researchers. Our universities and corporate R&D centres already produce outstanding talent, but the demand is set to surge.

Immediate Priority:

  1. Accurately Measure the AI Skills Gap. Establish a timely data-based view of how many AI professionals we need.

By 2030, to meet projected needs:

  1. Support Universities to expand their AI-oriented programs and align curricula with industry requirements.
  2. Increase Workforce Diversity, building on AI conversion courses while investing throughout the entire educational pipeline.
  3. Promote Non-University Pathways like apprenticeships, online courses, and self-directed study, ensuring flexible routes into AI careers.
  4. Create a Major AI Scholarship Programme on the scale of Rhodes or Fulbright, open to exceptional domestic and international students.
  5. Empower Ongoing Reskilling through “lifelong learning” programs, so both current and future workers can adapt to AI-driven change.

Attract Top Global AI Researchers:

  1. Establish a Government-Led ‘Headhunting’ Team to proactively recruit elite AI experts to British institutions, research labs, or startups.
  2. Expand Immigration Pathways for AI graduates from premier universities worldwide, minimizing visa barriers.
  3. Grow the Turing AI Fellowship, offering more opportunities to promising AI scholars, entrepreneurs, and innovators.

1.4 Fostering Trusted AI Development and Adoption via Regulation, Safety, and Assurance

Pro-innovation regulation is a British strength, encouraging new ideas while safeguarding public trust. Well-crafted rules make it easier to adopt AI products confidently and to foster innovation in emerging markets.

Key Actions:

  1. Continue Strengthening the AI Safety Institute (AISI) as a global leader in pre-deployment model evaluations, foundational safety, and societal resilience research.
  2. Overhaul Text & Data Mining Legislation, ensuring the UK’s regime is at least as flexible as the EU’s, addressing IP concerns that might hamper innovation.
  3. Increase Regulator Funding so they can build AI expertise and capacity.
  4. Instruct Regulators to Enable Safe AI Innovation, incorporating clear guidance on AI adoption into their strategic objectives.
  5. Streamline AI Sandboxes, particularly for high-growth, high-regulation sectors like autonomous vehicles and robotics.
  6. Measure Regulatory Impact on AI Innovation, requiring annual reporting on licensing times, guidance publication, and budget allocations.
  7. Invest in AI Assurance Tools, scaling programs that evaluate AI performance, reliability, and integrity—critical to earning public trust.
  8. Revisit the AI Institutional Landscape, exploring how the Alan Turing Institute could further drive frontier research, support government missions, and attract top talent.

2. Change Lives by Embracing AI

2.1 AI Adoption: Key to Achieving Government Missions

AI is pivotal to success in all five major government missions. From cutting the time teachers spend on administrative tasks to improving patient diagnosis and clinical decision-making, the right AI products can enhance outcomes for citizens and boost our economic competitiveness.

2.2 Government’s “Scan > Pilot > Scale” Strategy

Public-sector AI projects often remain siloed. A coherent, nimble, and forward-thinking approach is needed:

  1. Scan:

    • Assign an AI lead for each government mission, mapping AI use-cases to core challenges.
    • Establish a horizon-scanning unit that maintains up-to-date knowledge of AI capabilities.
    • Form partnerships with AI vendors to share public-sector needs early on.

  2. Pilot:

    • Adopt flexible frameworks for deciding whether to build solutions in-house, buy them, or run competitions.
    • Set up a rapid-prototyping arm with the technical skill to test new concepts.
    • Upgrade recruitment to attract more top AI talent into the civil service, compensating them competitively.
    • Make data accessible and streamline procurement so innovation can flourish.

  3. Scale:

    • Create a pathway for successful pilots to access additional resources and scale nationally.
    • Launch mission-focused, large-scale AI tenders for areas like health diagnostics or government administrative services.
    • Develop or procure an integrated AI technology stack to support nationwide digital services.
    • Standardize infrastructure, promote code reusability, and open-source code where possible.

2.3 Reinforcing Growth: Public and Private Sectors Together

By working in concert, the public and private sectors spur each other on:

  • Government Procurement: Use robust public demand to encourage private AI innovation.
  • Digital Government APIs: Mandate open interfaces so private companies can integrate with government systems, streamlining operations and promoting innovation.
  • AI Knowledge Hub: Provide a single platform to share best practices, frameworks, and real-world success stories.

2.4 Overcoming Private-Sector Barriers

AI adoption could add an extra £400 billion to the UK economy by 2030. To realize this potential, government should:

  • Tie AI to the New Industrial Strategy, tailoring adoption strategies for life sciences, finance, creative industries, and more.
  • Appoint AI Sector Champions to encourage and coordinate AI uptake within specific industries.
  • Spread AI’s Impact Nationwide, engaging SMEs, supply chains, and local communities, ensuring broad-based economic benefits.

3. Secure Our Future With Homegrown AI

By 2029, AI is on course to become a central pillar of global economic activity and national security. As large language models and agentic AI systems continue to evolve, steering this revolution—and sharing in its rewards—demands that we create and anchor AI champions in the UK.

The Next Frontier

Leading-edge AI continues to push boundaries in areas like:

  • Generative Models (images, video, and text)
  • AI for Science (protein folding, drug discovery)
  • Agentic AI (AI agents that reason, plan, and act toward goals)
  • Robotics and Embodied AI (real-world automation)

Each breakthrough brings enormous potential—and tremendous strategic influence. Nations commanding AI’s frontiers will guide how it develops and which values it encapsulates.

Building National Champions

The UK already hosts a remarkable concentration of AI talent and top-tier research labs, including DeepMind. Our time zone and position between the US and Europe make it easier for multinational teams to flourish here. Yet we need decisive, high-level coordination to compete effectively.

Key Action:

  1. Establish “UK Sovereign AI,” a new unit with the authority to unify and expand the country’s influence in frontier AI. Through deep collaboration with private companies, universities, and overseas partners, UK Sovereign AI will:
  • Channel Investment into next-generation AI startups, joint ventures, and strategic research areas.
  • Secure Access to Compute and fast-track data centre deployment via AI Growth Zones.
  • Leverage Public Data assets for pioneering UK-based research.
  • Attract Elite Teams from abroad, offering compelling reasons to choose the UK as their innovation base.
  • Influence Governance of emerging AI systems, ensuring Britain has an economic and strategic stake in how advanced AI evolves.

AI may be the defining technology of our time. Now is our chance to shape its progress, ensuring that British citizens, institutions, and businesses stand to gain from AI’s astonishing possibilities.

Conclusion

This Action Plan calls for immediate and long-term commitment. We must secure the foundational infrastructure, talent pipelines, and policy frameworks that pave the way for future breakthroughs. At the same time, bold public-private cooperation is needed to scale AI across the economy, deliver better services to citizens, and establish the UK as a global leader in the age of AI.

It’s time to act. Let’s build Britain’s place at the forefront of AI innovation—together.