AI Tools Reshape Grant Applications in UK Higher Education
The UK higher education sector is witnessing a profound shift as artificial intelligence tools increasingly influence the preparation and assessment of research grant proposals. UK Research and Innovation (UKRI), the primary public body allocating research funding, has issued clear guidance on the use of generative AI in applications and peer review processes. This policy emphasises maintaining integrity while recognising the potential efficiencies AI can bring to a system under strain from rising application volumes.
Researchers at institutions across the country, from the University of Sheffield to leading Russell Group universities, are exploring how these technologies might streamline workflows. However, the rapid adoption raises questions about fairness, originality, and the core principles of peer review that have long underpinned funding decisions.
UKRI's Policy Framework for Generative AI
UKRI's policy on generative artificial intelligence in application preparation and assessment, published in 2024, provides a structured approach for applicants, reviewers, and assessors. The guidance permits the use of AI tools for drafting and editing but stresses that applicants remain fully responsible for the content submitted. Reviewers are explicitly prohibited from inputting confidential proposal details into AI systems to generate assessments.
This balanced stance reflects broader discussions within the Research Funders Policy Group, which includes representatives from major UK funders. The policy aims to harness AI's benefits, such as improved clarity in proposals, while mitigating risks like hallucinated references or biased outputs. Universities are now incorporating these guidelines into their internal research support services, offering training sessions for academics navigating the new landscape.
Surge in Applications Prompts AI Experimentation
UKRI has reported a significant increase in grant applications in recent years, with award rates halving over the past seven years despite an 80% rise in submissions. To address this pressure on the peer review system, the organisation has opened access to anonymised data from up to 2,000 proposals. Researchers, including data scientist Mike Thelwall at the University of Sheffield, are investigating whether generative AI can assist in initial screening or summarisation tasks.
Funded through the UK Metascience Unit, these experiments represent a proactive response to the administrative burden. Early findings suggest AI could help identify promising proposals more efficiently, allowing human reviewers to focus on nuanced evaluation. Institutions such as the University of Manchester and Imperial College London are monitoring these developments closely, considering how similar tools might be adapted for their own internal funding competitions.
Concerns Over AI-Generated Proposals
While efficiency gains are appealing, many academics express concern that AI agents capable of autonomously generating complete grant applications could overwhelm the system. Reports indicate that funders are already seeing a rise in AI-assisted submissions, prompting fears of reduced originality and increased similarity across proposals.
Experts highlight the risk that sophisticated AI tools might produce technically polished but substantively shallow applications, potentially disadvantaging researchers who rely on traditional methods. This has led to calls for enhanced detection mechanisms and clearer disclosure requirements. The University of Bath's PVC for Research has publicly discussed these challenges, noting the need for funders to adapt policies rapidly.
Photo by National Cancer Institute on Unsplash
Responsible AI Use in Funding Decisions
A new handbook from the Research on Research Institute (RoRI), developed in partnership with international funders including UKRI, offers practical guidance on responsible AI deployment in research funding. The publication stresses transparency, equity, and ongoing evaluation of AI tools to avoid unintended biases.
UKRI's own experiments align with these principles, emphasising human oversight at every stage. Training programmes for peer reviewers now include modules on AI literacy, ensuring assessors can critically evaluate AI-influenced proposals. This collaborative approach involves organisations such as the Wellcome Trust and the Royal Academy of Engineering.
Impact on Early-Career Researchers
Early-career academics in UK higher education stand to benefit from AI tools that assist with proposal structure and language refinement, particularly those for whom English is an additional language. However, there are worries that over-reliance could undermine the development of independent research skills.
Universities are responding by embedding AI ethics into doctoral training programmes. Centres for Doctoral Training supported by UKRI, including those focused on AI and sustainability, are incorporating responsible use modules. This ensures the next generation of researchers is equipped to navigate the evolving funding environment.
Future Outlook for Grant Awarding Processes
Looking ahead, AI is likely to play an expanding role in UK research funding, from initial triage to post-award monitoring. UKRI continues to invest in metascience research examining how AI transforms the research landscape. Pilot programmes testing AI-assisted review are expected to expand, subject to rigorous evaluation of fairness and effectiveness.
Stakeholders across the sector, including Universities UK and individual institutions, advocate for ongoing dialogue to shape policies that preserve the integrity of peer review while embracing technological advances. The goal remains a funding system that is efficient, equitable, and supportive of high-quality research.
Stakeholder Perspectives Across UK Universities
Views among UK academics are divided. Some welcome AI as a tool to level the playing field and reduce administrative load, while others fear it could erode the human judgment central to assessing research potential. Vice-chancellors and research directors at institutions like the University of Edinburgh and University College London emphasise the importance of maintaining trust in the system.
Student and early-career researcher representatives have called for greater transparency in how AI tools are used during assessment. This feedback is informing UKRI's iterative policy updates, ensuring the voices of those most affected are heard.
Photo by Daria Nepriakhina 🇺🇦 on Unsplash
Practical Steps for Researchers and Institutions
UK universities are advised to update their research support services to include guidance on AI use in grant writing. Workshops on prompt engineering, bias detection, and ethical disclosure are becoming standard. Researchers are encouraged to document AI assistance in proposals where required by funders.
Institutions can also contribute to sector-wide learning by sharing anonymised data and participating in metascience initiatives. This collaborative spirit is essential for developing best practices that benefit the entire UK higher education community.
