Britain's year of AI delivery: better care, smarter schools, stronger growth

The UK's AI Opportunities Action Plan: One Year In, Built For Delivery

One year ago, the plan was simple: use AI to grow the economy, improve public services, and help people build the skills to thrive. Over the last twelve months, that moved from slide decks to real-world delivery.

This update is for people in government who need the signal, not the noise. What got built, what's working, and what to do next to put it to work in your department or region.

What's been delivered so far

• People first: Over 1 million free AI upskilling courses delivered as the first step toward 10 million workers by 2030. New scholarships and apprenticeships are live.
• Public services: AI is assisting one-third of NHS chest X-rays (2.4 million scans), pilots of AI tutoring are launching, and AI tools are speeding up planning decisions.
• Economy: 5 AI Growth Zones designated, unlocking £28.2 billion in investment and creating 15,000 jobs. UK compute capacity is expanding at pace, including Isambard-AI in Bristol.

1) Lay the foundations to enable AI

Compute and infrastructure

• AI Growth Zones: 5 zones across Great Britain (including 2 in Wales and one in Scotland) with reforms to planning and energy access, plus £5 million per zone for local adoption. A dedicated Delivery Unit will accelerate electricity, planning, and offtake so data centres get built where they're needed.
• Public compute expansion: Isambard-AI was switched on in July 2025. Up to £250 million has been earmarked to scale cloud capacity for the AI Research Resource. Supercomputer capacity at Cambridge (home to DAWN) is increasing sixfold by Spring 2026.
• Secure and sustainable: Zones are located near renewable and new low-carbon energy, including SMRs in North Wales and a micro-grid in Lanarkshire. DSIT, AISI and NCSC have launched a research programme to make AI computing systems secure by design.

Data assets that matter

• National Data Library: Backed by over £100 million to curate high-value public data, publish AI-ready guidance, and run open calls so researchers and industry can propose high-impact datasets.
• Health data at scale: The new Health Data Research Service is creating a secure single access point to regional and national datasets, with up to £600 million in joint government and Wellcome funding. Leadership is in place, with CEO appointed in January 2026 and Chair in November 2025.
• Education datasets: Work is underway on an AI Education Content Store to improve accuracy for curriculum and assessment use cases.
• Data + compute pairing: Isambard-AI will link to a new "British Library for the AI age" data facility. The next national supercomputer (backed by £750 million) will sit alongside the International Data Facility at the Edinburgh Parallel Computing Centre for secure, large-scale research.
• Private and research data: The Creative Content Exchange is piloting with 12 cultural institutions to enable large-scale licensing of digitised assets. The Data (Use and Access) Act 2025 enables Smart Data schemes, backed by £36 million, so consumers and SMEs can share data securely to access more relevant products and services.

Skills, talent, and visas

• Upskilling at scale: Over 1 million free AI courses delivered via the AI Skills Boost platform. Goal: 10 million workers with core AI skills by 2030.
• TechFirst investment: £187 million committed. Highlights include £7.8 million for the TechLocal AI Professional Degree and Traineeship Accelerator and the launch of TechGrad undergraduate scholarships with paid work placements (350 students in 2026, scaling to 900 a year).
• Spärck AI scholarships: Full tuition and stipends for hundreds of master's students across 9 universities to build the UK's next wave of AI specialists.
• Global talent: A new AI sourcing function has scaled AISI to 100+ technical staff, drawing alumni from leading labs. The Global Talent Taskforce has doubled resourcing and strengthened concierge support. Visa fee reimbursements and strengthened routes (High Potential Individual, Innovator Founder, Global Talent) make it easier for top talent to relocate and build here.
• Future of work: A cross-government unit is mapping AI's labour market effects, advising on timely policy, and ensuring people transition into better jobs with minimal disruption to communities.

Safety, assurance, and regulation

• AI Security Institute (AISI): Backed by £240 million. 100+ researchers, 30 frontier models tested, and a growing evidence base on capabilities and risks. Outputs include the Frontier AI Trends Report, 10 NeurIPS 2025 papers, a Science study on AI-driven persuasion, and the international Alignment Project. AISI chairs the International Network for AI Measurement, Evaluation and Science.
• Assurance market: The Trusted Third-Party AI Assurance Roadmap, an £11 million AI Assurance Innovation Fund (first round opens Spring 2026), and the new Centre for AI Measurement at the National Physical Laboratory are building testing, verification, and validation capabilities the UK can trust.
• Regulatory innovation: Work is underway to stand up the cross-economy AI Growth Lab sandbox. Regulators are moving: MHRA's Airlock phase 2 includes clinical AI assistants; the FCA's Supercharged Sandbox (with Nvidia) improves access to data, technical expertise, and regulatory support; and new funds (£3.6 million AI Capability Fund and £8.9 million Regulators' Pioneer Fund) are building regulator capacity. Regulators will publish plans on enabling safe AI adoption and report annually.

2) Change lives by embracing AI

Public service delivery, scaled

• NHS diagnostics: One-third of chest X-rays are now AI-assisted, speeding up decisions and improving care. Stroke treatment and prostate cancer diagnosis times are falling in early deployments.
• Planning: Extract converts planning documents into structured data, cutting time and errors. A new decision-support tool targets a 50% reduction in processing time for straightforward applications, with regional trials by May 2026 and national scale-up in 2027.
• Local government productivity: Minute, an AI scribe for meetings, now supports 1,000 officials across 22 local authorities and is moving to a live service model.
• Education: Trials of AI tutoring tools are being launched to improve attainment and reduce teacher workload, with the potential to support up to 450,000 pupils on free school meals in Years 9-11.

• Procurement that moves faster: AI Accelerator Tenders target regional-scale products in about six months. The Incubator for AI (i.AI) runs a single front door for suppliers. From January 2026, a clear AI Commercial Strategy prioritises buying from the market and challenge-led procurement.

Private sector adoption with sector leadership

• Modern Industrial Strategy: Eight priority sectors now have AI Sector Champions, with funding to remove blockers and scale adoption. That includes £150 million for six transformative AI and technology programmes and £40 million for Robotics Adoption Hubs and an expanded Made Smarter programme.
• BridgeAI expansion: Innovate UK is scaling BridgeAI across priority sectors to give firms practical support, funding, and expert guidance to adopt AI before the end of this Parliament.

3) Secure our future with homegrown AI

• Sovereign AI Unit (year one): Allocated sovereign compute to researchers and high-growth start-ups, including support for Cambridge's MACE model for materials discovery. Invested £8 million in the OpenBind structural dataset for AI-enabled drug discovery. Scaled the Pillar VC and ARIA Encode Fellowship to bring more top AI founders and researchers to the UK.
• Next phase (April 2026): Backed by up to £500 million and chaired by James Wise, the unit will invest to help UK AI companies become world-leading in critical parts of the value chain.

What this means for government teams

AI is now baked into the UK's growth plan and public service reform. Your task is to translate national architecture into local delivery: plug into zones, datasets, and compute; adopt proven tools; and build skills across teams.

Practical next steps for departments, agencies, and councils

• Pick 2-3 high-impact use cases you can stand up in 90 days (planning triage, document extraction, meeting notes, citizen queries). Use existing tools where possible.
• Run assurance early: Engage your risk, legal, and data protection teams from day one. Where relevant, prepare to use the AI Growth Lab sandbox or regulator sandboxes (e.g., FCA, MHRA).
• Do a quick data inventory: Identify AI-ready datasets, gaps, and owners. Align with the National Data Library's guidance to make priority datasets AI-ready.
• Use procurement that fits the pace: Start with AI Accelerator Tenders or challenge-led calls. Keep pilots small, then scale to region-wide deployments.
• Skill up your people: Set a minimum AI skills baseline for policy, delivery, and ops teams using free courses already available. For structured learning by role, see Complete AI Training: Courses by Job.
• Connect to Growth Zones: If you depend on data centre capacity, engage the AI Growth Zone Delivery Unit early to resolve electricity, planning, and offtake questions.

What's next in 2026

• Move AI Growth Zones from designation to live data centre capacity with fast-tracked electricity and planning.
• Scale proven public service tools nationally: diagnostics, planning support, digital assistants, and AI tutoring.
• Operationalise the Sovereign AI Unit's next phase to back UK firms across the AI value chain.
• Open the AI Assurance Innovation Fund and continue building a trusted, credible assurance market.
• Publish regulator plans for safe AI adoption and start annual reporting.

The bottom line

The foundation is set: compute, data, skills, safety, and early wins in public services. The job now is disciplined execution-pick high-value use cases, ship quickly, measure outcomes, and scale what works.