After months of preparation, consultation and analysis, the Research and Development Society is excited to present the final version of its report into Forging the UK's AI Future: A Blueprint for UK AI Policy. As the country steps into a new technological era, it is vital that it adopts a clear and integrated strategy that presents an ambitious vision of the future. In this report we lay out some aspects of such a vision and our recommendations on how to get there.
Executive Summary
Purpose. This is the first output from the RDS "Big Tent" coalition under its Manifesto for National Growth. It argues that UK industrial strategy lacks a defined model of growth, and that the missing link is between macro-economic policy levers and the meso-economic vectors affected by the levers. AI matters disproportionately as it is a pervasive technology that will restructure nearly every market space. The report covers horizontal AI plus four market spaces: life sciences, media & entertainment, energy & transport, and defence & security.
A seven-part vision for 2030. Regulation as an export (other countries building on UK frameworks); at least five sovereign AI champions scaling in the UK rather than abroad; dominance of the "hybrid layer" between foundation models and domain-specific technologies; world leading preventative healthcare via the NHS; reversal of rising grid-constraint costs; a trusted media environment; and Europe's cyber security hub.
Positioning. The UK should not chase semiconductor manufacturing or frontier model development, where it cannot compete with the US and China. Fabs should be treated as security infrastructure, not growth ventures. Effort should go up the product stack into hybrid technologies. Sector-specific: full-breadth ambition in life sciences; content creation and trusted-institution/verification tools in media, but not distribution infrastructure; grid and distribution management in energy; cyber and aerospace (GCAP) in defence, with deliberate NATO niche specialisation.
Sovereignty. The report criticises the absence of a working definition and offers its Technology Sovereignty Index — seven axes covering expertise, legal authority, taxation, strategic decision-making, economic embeddedness, infrastructure, and investment sources. Worked examples contrast DeepMind, Arm, and Wayve, and show sovereignty eroding after ownership or leadership changes. Sovereignty priorities should be weighted differently by sector rather than pursued uniformly.
Diagnosis by policy lever (ordered by size of gap):
- Regulation — simultaneously permissive on fundamental development and obstructive on deployment in the UK's strongest sectors. The principles-based approach offloads democratic trade-offs onto sector regulators and leaves training-data policy uncovered.
- Signalling — sovereignty rhetoric contradicted by procurement practice; research bodies over-securitised (Alan Turing Institute re-tasking); negative labour-market consequences under-acknowledged.
- Global drivers — trade agreements largely silent on AI; over-alignment with the US; defence thinking treats information disorder as foreign interference rather than structural.
- Procurement — US firms (Microsoft, Palantir) dominate; exemptions have damaged trust; vendor lock-in risk.
- Funding — a relative strength (3rd globally for private AI investment), but public allocation skews to early-stage and to US suppliers.
- Infrastructure — physical infrastructure adequate for mid-range scenarios, but data-centre demand could reach ~16% of UK electricity by 2035; the National Data Library is largely unusable due to poor labelling and meta-data.
- Talent — skills definitions are sound but assume unchanged market structures; the AI Skills Hub is a poorly organised catalogue built around vendor tech stacks; entry-level jobs are contracting.Taxation — least impactful, though NIC relief for new entrants and agent taxation are floated; Patent Box is useless for pre-profit firms.
Conclusion. The UK has the infrastructure, talent, and capital, but sovereignty talk must translate into accepted trade-offs — lower US inward investment, riskier bets on domestic suppliers — alongside a clearer vision, principles with enough clarity for regulators to act, and in-house digital capacity rather than outsourced dependency.