Fellowships launched to explore how AI could change the way scientists drive new discoveries

New government-funded fellowships exploring how AI could change the way scientists drive future discoveries are now open for applications, Science Minister Lord Vallance has announced today (Thursday 13 February).

Metascience – the study of how science works – examines research practices, funding models, and how institutions operate to improve how science and research is conducted, and discoveries are made and applied. By understanding what makes scientific research more effective, metascience helps drive research breakthroughs faster and with greater impact – boosting economic growth and prosperity to drive our Plan for Change.

The AI Metascience Fellowship Programme will fund research into key questions, including how AI is reshaping the research landscape and both changing and supporting the daily work of scientists. It will explore ethical concerns such as biases in AI-driven research and transparency in AI-generated discoveries and economic impacts like shifts in research jobs and funding priorities.

The new fellowship builds on the momentum of the recent AI Action Summit, as global leaders work to ensure AI's development benefits society and be rolled out across society in the public interest.

AI is already revolutionising research with DeepMind's AlphaFold accelerating drug discovery, while AI-powered lab robotics are automating complex experiments and machine learning is transforming how scientists analyse vast datasets.

The programme will also examine how governments and businesses should respond, from ensuring AI-driven science remains rigorous and delivers reliable outcomes to supporting researchers to maximise their creative potential and spend less time on mundane tasks.

Funding will go towards researchers to apply their expertise in examining the technology's broader effects on research. The £4 million UK programme will run alongside a US-based cohort funded by the Alfred P. Sloan Foundation, creating a transatlantic research effort to examine AI's impact on science. Fellows from both countries will attend a fully funded summer school, strengthening international collaboration and knowledge exchange.

Applications are especially encouraged from projects exploring the impact of AI on research jobs and skills, how it affects the speed of scientific progress, and the challenges of ensuring AI-driven research remains reliable and explainable.

Science Minister, Lord Vallance said:

AI presents new opportunities in a range of sectors, and if researchers can demonstrate its potential to increase transparency, robustness and trust in science then this could pave the way to freeing them up from mundane paperwork tasks while driving growth.

Supporting researchers to explore how AI can change the way we conduct research and through our joint support with Open Philanthropy for 23 projects exploring wider research and innovation, we will build a better understanding of what works in research - maximising impact, driving discoveries and improving lives.

In addition to the Fellowship, Department for Science, Innovation and Technology (DSIT) and UK Research and Innovation (UKRI) have awarded £4.8 million in funding for 23 new research projects, which will tackle key questions about how to improve research and innovation, including AI's impact on science, research integrity, and new models for funding and publishing research.

It follows a funding call launched last year and includes £1.8 million in co-funding from Open Philanthropy, a US-based foundation.

Among the winning projects:

• University of Sheffield: Assessing whether large language models – like ChatGPT – can reliably review academic work and contribute to the UK's Research Excellence Framework and journal peer review
• University of Bath: Partnering with Sage Publishing and the Royal Society to test a two-stage peer review process, designed to increase trust in academic findings
• University College London (UCL): Working with Google DeepMind and the UN Development Programme (UNDP) to explore how AI-driven research can be applied to global challenges, including sustainability and healthcare

Notes to editors

List of the Metascience grant winners.

AI Peer: Large language models and academic peer review outcomes
Michael Thelwall, University of Sheffield.

Analysing the Reliability of Quantitative Impact Evaluations (ARGIE)
Jack Blumenau, University College London.

Assessing compliance with the FAIR Guiding Principles: a systematic evidence map of data availability in metabolomics research
Matt Spick, University of Surrey.

Big Science Beyond Science: The Innovation Impact of Research Infrastructure Procurement
Riccardo Crescenzi, LSE.

Commercialising Deep Tech: Understanding Frictions to University Invention Disclosure
Ramana Nanda, Imperial College London.

Cultural Traction: Embedding research culture strategy
S Martin Holbraad, University College London.

Evaluating the Development and Impact of AI-Assisted Integrity Assessment of Randomised Trials in Evidence Syntheses
Alison Avenell, University of Aberdeen.

Everything we (think we) know about Narrative CVs
Liz Simmonds, University of Cambridge.

Financial structures for enabling innovator participation and success: experimental evidence from challenge prizes
Vidal Kumar, Nesta.

Fostering a Dynamic Academic Ecosystem: Innovative Platforms and Methodologies for Econometrics
Martin Weidner, University of Oxford.

Making Replications Count: Identifying Barriers and Enhancing Impact with Innovative Dissemination Tools
Lukas Wallrich, Birkbeck, University of London.

Mapping impact pathways: improving our understanding of what mechanisms work in research translation
Alexandra Pollitt, King's College London.

Metascience, research funding and policy priorities
Annette Boaz, King's College London.

People or Projects (PoP)? Investigating different research funding styles
Ohid Yaqub, University of Sussex.

PRIME: Peer Review Improvement for Minimizing Bias in Evaluation
Katherine Button, University of Bath.

Providing empirical evidence to support greater equality, diversity, and inclusion (EDI) in research funding
Philip Clarke, University of Oxford.

Public value mapping for AI
Jack Stilgoe, University College London.

Research Software Engineer Metascience
Heather Packer, University of Southampton.

Sharing Code for Medical Research: An Audit Tool and Pilot at The BMJ
Nicholas DeVito, University of Oxford.

Supporting Research and Researchers through the deployment of Digital Notebooks: A framework for implementation and impact
Andrew Stewart, University of Manchester.

Transparent and Reproducible Science in the 21st Century: Unlocking the Benefits of Open Source Code
Albert Bravo-Biosca, Nesta.

Understanding Scientific Prizes - Structure, Evolution and Impact
Ching Jin, University of Warwick.

Working together or writing together?
Steven Wooding, University of Cambridge.