UK’s new Metascience Unit is testing AI tools to overhaul research assessment and boost policy evidence.
The United Kingdom has created a dedicated Metascience Unit within the Department for Business, Innovation, Science and Trade (BIST) and UK Research and Innovation (UKRI). Its remit is to apply the scientific method to the very systems that fund, organise and evaluate research, treating the research ecosystem itself as an object of rigorous study.
By doing so, the unit hopes to generate evidence that can steer policy, improve funding efficiency and ultimately accelerate scientific breakthroughs.
Operating at the intersection of economics, sociology, data science and the history of science, the Unit brings together scholars who use social-scientific methods—including experiments, qualitative analysis and large-scale data work—to interrogate how science works in practice.
It also cultivates a community of practice that links researchers, funders, universities and private enterprises around the shared goal of smarter science policy.
Three pillars of the Metascience Unit’s work
First, the Unit administers multi-million-pound competitive grants that nurture the UK’s metascience community and promote applied research aligned with policy needs.
These programmes also foster international collaborations with peer funding bodies, extending the impact of UK-led investigations beyond national borders.
Second, it designs and runs a portfolio of internal experiments to test what works in UKRI’s funding processes. This includes large-scale randomised controlled trials, analyses of historic grant data, shadow experiments and policy pilots, all aimed at uncovering the most effective levers for improving research allocation.
Third, the Unit disseminates its findings to BIST, UKRI and the broader funding landscape, ensuring that decision-makers have access to the latest metascience evidence. By maintaining a presence inside central government, the Unit keeps its agenda aligned with national research and development strategies.
AI-centred research agenda for 2026-2028
Redesigning assessment and decision-making with AI
The Unit’s flagship theme, “Research assessment and decision-making in an age of artificial intelligence” tackles the surge in grant applications that strains traditional peer-review mechanisms. Projects under this theme explore distributed peer-review models that scale human expertise, while simultaneously piloting AI-supported tools for evaluating grant proposals, journal submissions and portfolio performance. A key aim is to upgrade UKRI’s internal data infrastructure so that AI can provide real-time, evidence-based guidance to funding committees.
Mapping the impact of AI on scientific practice
Another strand, “Understanding how AI is changing science,” launches a National AI for Research Survey to capture how researchers across disciplines are adopting machine-learning techniques. The Unit also backs dedicated fellowships and grant schemes that enable scholars to investigate AI’s influence on experimental design, data sharing and collaborative networks. These efforts seek to create a detailed picture of AI’s ripple effects throughout the research ecosystem.
Building next-generation scientometric indicators
In the “Advancing scientometrics” theme, the Unit is developing and validating novel metrics that go beyond citation counts. Indicators of noveltyrigour and societal impact are being tested for reliability and bias, with an eye toward assembling a “basket of indicators” that can serve as a new quantitative vocabulary for research quality. The programme also encourages critical reflection on the limits of any single metric, emphasizing a balanced, modular approach.
Strategic significance and international cooperation
Public investment in research and innovation runs into the billions of pounds each year, yet historically little attention has been paid to optimising the way that money is spent. The Metascience Unit argues that improving the quality of funding mechanisms is as vital as the volume of money, because it directly influences policy outcomes, career pathways, economic growth and the ability to tackle future challenges.
Beyond national work, the Unit is collaborating with the U.S. National Science Foundation under the “Kyoto Vision for a Golden Age of Science.” This partnership will explore joint metrics for productivity and impact, share best practices in research review, and lay the groundwork for an international network of metascience centres. While the Unit generates evidence, it explicitly states that setting policy is not its role; instead, it provides the data that policymakers can use to make informed decisions about AI-driven evaluation.
