Aim
This funding opportunity aims to accelerate the generation of evidence on how we can improve the efficiency, effectiveness, and inclusivity of the research and development (R&D) ecosystem. For this round, we are interested in how the adoption of AI is changing the research landscape , how to optimally design and lead research institutions; and how to measure and understand scientific progress at scale.
Scope
Metascience, a rapidly expanding research field, draws on a wide range of disciplinary expertise to understand how research is conducted, funded and supported. It also provides evidence for how these practices can be enhanced or improved. For a deeper understanding of what metascience means to UKRI, please see the UK Metascience Unit’s report: A year in metascience (2025) – GOV.UK.
The Metascience Research Grants Programme, a collaboration between UKRI and Coefficient Giving, supports innovative and ambitious metascience research projects. These projects use scientific methods to deepen our understanding of how different incentives, institutional structures, and funding practices within the R&D system influence scientific research outputs and career outcomes.
This funding opportunity will support empirical and/or theoretical research that is focused on generating actionable insights for decision makers, including those in government, funding bodies, and research organisations.
In this funding opportunity, we are focussing on three themes to build our metascience portfolio. Applications should fit under one of the following themes.
Science of AI for Science:
How the adoption of AI is changing the research landscape, how this helps and/or hinders scientific progress, and how governments, industry and funding organisations should respond.
Effective design and leadership of research organisations
This includes empirical comparison of institutional models, the drivers of programme manager and research performance, the application of evidence from management and behavioural science to improve organisational structures and practices in research environments, and the effectiveness of interventions to support inclusive, high-performing research cultures.
Scientometric approaches to understanding research excellence, efficiency, and equity
This includes the development, validation, and generalisable use of metrics and indicators to assess research quality, influence, and impact, the development or application of indicators to advance the curation and synthesis of science at scale, and the behavioural consequences of metric use in research evaluation and funding decisions.
We will not fund applications that do not fit under one of these three themes. In your application, you should clearly state the theme your proposal fits within, alongside providing a clear justification.
The funders strongly welcome projects involving collaborations between researchers and organisations (for example research funders, research organisations, charities, think-tanks, and journals) interested in implementing findings or approaches from the proposed research in their practices.
Science of AI for Science
As an emerging area, it is our experience from other funding opportunities that AI for science requires further guidance to ensure common understanding.
We define AI broadly as ‘software which learns by example’, including generative AI and machine learning, and applications of these in hardware, for instance, self-driving laboratories. We define ‘AI for science’ as the application of AI in scientific research itself (including social science) and in activities undertaken within a research ecosystem, for instance, peer review or research portfolio evaluation.
This funding opportunity aims to fund projects that contribute to the embryonic ‘science of AI for science’, or ‘AI metascience’. These are projects that will generate broad understanding and evaluations of the use of AI and its impacts that is relevant across multiple scientific fields and contexts.
We will reject projects focussed primarily on the application of AI in industrial settings like clinical medicine, law or fintech. We will also reject proposals focussed on conducting frontier computer science research (i.e. the ‘science of AI’, as opposed to ‘AI metascience’), or on general AI ethics, security, safety and society-related topics.
This is not because these are not important, but because they are covered much more substantially in other programmes funded by UKRI.
Duration
The duration of this award is between six months and 24 months.
Funding available
The FEC of your project can be up to £250,000 if all organisations are UK-based and eligible for funding. The FEC of your project can be up to £350,000 if you have an international partner.
UKRI will fund 80% of the FEC.
