Science · Innovation · AI
Improving how science is funded
As technical partner to the UK Metascience Unit, RAND Europe is helping to build evidence on how research is funded, reviewed and organised – and how AI may change it.
From the chapter “Advancing the UK’s metascience goals” · Report pp. 16–19
Why how science is funded matters
Government estimates cited in the report indicate that, on average, £1 of civil public R&D investment generates £8 in net economic benefits for the UK over the long term. Yet the way R&D is organised, funded and incentivised may hamper progress, through inefficient funding allocation, potential bias and conservatism in peer review, and barriers to commercialisation.
In autumn 2025, the Department for Science, Innovation & Technology set out its R&D budget for the four years from April 2026, including around £9.2bn in 2026–27 for UK Research and Innovation (UKRI), and announced ‘a new, more agile approach’ to managing R&D funding.
A technical partner to the Metascience Unit
The UK Metascience Unit was established in 2024 by the UK government alongside UKRI, with RAND Europe as its technical partner. It is the first body of its type to be formally based within government, created because limited attention had been paid to how science itself is funded and managed.
Working alongside the Unit, RAND Europe has helped to better understand distributed peer-review processes, generated evidence on the financial health and sustainability of UK university R&D, provided insight on strengthening international collaboration, and is contributing to the UK government’s AI for Science Strategy. RAND Europe is also working with the Unit to run a global competition to identify AI-driven indicators of scientific novelty, and a major observational study is under way on what drives success in interdisciplinary research through UKRI’s Cross Research Council Responsive Mode scheme.
Peer review under scrutiny
Peer review shapes nearly all of UKRI’s £8.8 billion annual funding portfolio and safeguards scientific quality and accountability. But, as the independent Tickell Review highlighted, it is often criticised as slow, bureaucratic and resource-intensive. A key focus of the Unit’s year has been UKRI’s grant allocation processes, responding to the Tickell Review and UKRI’s own Review of Peer Review.
The opportunities and risks of AI in research
AI is seen as a way to increase the speed and productivity of research. The report also records concerns about reliability, reproducibility and transparency; ethical risks relating to privacy, bias and access; and the potential long-term degradation of some research skills. In grant-making, AI could help match proposals to priorities, but a lack of transparency could perpetuate biases such as favouring past success over novelty. RAND Europe is working with the Unit to gather evidence and perspectives on the use of AI in research.
“Solving the societal challenges of the 21st century crucially depends on scientific progress. Scientific discoveries, in turn, need to be enabled by a system that effectively supports researchers and their research. Our work on metascience helps build a robust body of evidence on questions such as how best to organise peer review processes, methodological innovations in research evaluation, and scientific novelty itself. Working in direct cooperation with UKRI’s Metascience Unit, we are well positioned to achieve immediate practical impact.”
Key findings and implications
- Adding funding to an overstretched R&D system risks failing to address the issues that limit its impact.
- Peer review is central to research integrity but is often criticised as slow, bureaucratic and resource-intensive.
- AI could increase research productivity, but raises concerns about reliability, transparency, bias and skills.
- Sharing learning about AI before bad practice becomes embedded across the system is critical.
Source
RAND Europe, Spotlight 2025–2026 (CP-A1150-6), printed pp. 16–19 (PDF pages 18–21). Text adapted for the web; quotations reproduced as published. Statements reflect the period of publication.
Image credits: Illustration: Dàlia Adillon (printed pp. 16–17).
Related research
Health
Making innovation part of healthcare
Exploring how research and innovation can support safer care, better access and a more sustainable health system.
Futures
Preparing for possible futures
Using foresight methods to help policymakers explore uncertainty, test decisions and prepare for emerging challenges.