UK research funders have taken a significant step forward in embracing generative artificial intelligence for the administrative processing of grant applications, while firmly maintaining human oversight for all final funding decisions. This clarification, announced at the Association for Research Managers and Administrators conference in June 2026, represents a pragmatic evolution in policy for major organisations including UK Research and Innovation and the Wellcome Trust.
From Caution to Clarification in AI Policy
Previously, UK funders operated under a joint statement from the Research Funders Policy Group that largely restricted the use of generative AI tools in both application preparation and assessment. UKRI’s existing policy, for instance, prohibited assessors from employing such tools for reviewing or scoring applications. Wellcome similarly emphasised that it does not use generative AI to assess quality or aid funding decisions.
The updated position softens this stance for processing activities. Funders may now utilise generative AI to handle administrative tasks associated with grant applications, such as formatting, summarising, or managing workflows, provided that entire applications are never fed into publicly available AI tools and that confidentiality is preserved. Final decisions on funding will continue to rest exclusively with human reviewers and panels.
Key Elements of the Revised Approach
The shift addresses the growing volume of AI-assisted applications reaching funders. Representatives noted that the Research Funders Policy Group, which includes UKRI, Wellcome and six other major bodies, will soon publish an updated joint statement reflecting these changes. The emphasis remains on transparency, with applicants still required to declare any use of generative AI in preparing their submissions.
Human oversight is the non-negotiable safeguard. While AI can streamline repetitive administrative work, the intellectual and evaluative core of the process stays firmly in human hands. This balances efficiency gains with the need to uphold research integrity and fairness across UK universities.
Implications for UK Universities and Researchers
For academics and research support teams at institutions across the United Kingdom, the policy update offers practical relief from administrative burdens. Grant writing is notoriously time-intensive; AI tools can assist with drafting structures, checking compliance, or generating summaries, freeing researchers to focus on the scientific content.
University research offices are likely to update their internal guidance and training in response. Staff responsible for pre-award support will need to ensure that any AI-assisted processing aligns with funder requirements, particularly around data protection and intellectual property. Early adopters at Russell Group universities may pilot compliant workflows to share best practice with the wider sector.
Benefits for Efficiency and Accessibility
Proponents highlight that the measured adoption of generative AI could level the playing field for smaller institutions and early-career researchers who often lack extensive administrative support. By automating routine processing steps, funders can handle higher volumes without compromising quality, potentially accelerating decision timelines.
UK higher education stands to gain from faster turnaround on grant outcomes, enabling quicker mobilisation of research projects in priority areas such as climate science, health innovation and digital technologies.
Safeguards and Remaining Restrictions
Despite the green light for processing, strict boundaries persist. Applicants must not submit wholly AI-generated applications, and reviewers are still barred from using generative AI in assessment activities. Funders have reiterated that confidentiality of proposals must be protected at all times.
These safeguards reflect ongoing concerns about originality, bias and the risk of over-reliance on AI outputs. Training and clear institutional policies will be essential to help researchers navigate the new landscape responsibly.
Stakeholder Perspectives Across the Sector
Research managers have welcomed the clarity, noting that it reflects the reality of how tools are already being used in practice. University leaders see opportunities to streamline operations while preserving the human judgement that underpins peer review.
Some early-career researchers express cautious optimism, viewing the policy as recognition that AI can be a supportive tool rather than a threat. Others stress the importance of maintaining the distinctive voice and originality that distinguishes strong proposals.
Comparative Context with International Funders
The UK approach aligns with evolving stances elsewhere. Australia’s National Health and Medical Research Council has similarly permitted generative AI as a support tool under human oversight. European Commission guidelines also stress transparency and accountability when research funding organisations deploy AI in internal processes.
By updating its position, the UK maintains competitiveness in attracting and supporting world-class research talent.
Photo by Miguel A Amutio on Unsplash
Future Outlook for AI in UK Research Funding
Further refinements are expected as experience with the updated policy accumulates. Funders have signalled ongoing dialogue with the research community to monitor impacts and address emerging challenges.
Over the coming years, integration of AI into grant administration is likely to deepen, provided human oversight remains central. This measured evolution could serve as a model for other countries balancing innovation with integrity in research funding systems.
Practical Advice for Academics and Institutions
Researchers should familiarise themselves with the forthcoming updated joint statement and ensure any use of generative AI is declared where required. University research support teams are advised to develop clear internal protocols and offer training sessions on compliant AI use.
Institutions may also consider investing in secure, institution-approved AI tools to minimise risks associated with public platforms. Collaboration across the sector, perhaps through groups such as the Association for Research Managers and Administrators, will help disseminate effective practices.
