Bristol-led report highlights potential of AI in assessing Higher Education research

By Pippa Largey, Second Year, English and Philosophy

Research Excellence Frameworks (REF) are held every six to seven years to assess the research of Higher Education Institutions (HEI). They help determine how £2 billion of public research funding is allocated to universities.

In the 2021 REF the University of Bristol was ranked 5th for research across UK universities. In this year's Times University Ranking it received a 66.2% research quality score, having placed 10th overall.

The next REF will be held in 2029.Richard Watermeyer, Professor of Higher Education at the University of Bristol, has led a report into ‘the potential of generative AI for REF2029.’

According the report, Generative AI has been used to varying extents and in different ways by HEIs in preparation for REF submissions.

Some universities have been using AI to gather evidence of their research impact in the wider world. Others have been using it to review, assess and score their research.

Epigram looks inside the new Temple Quarter Enterprise Campus

Accessible directly from Temple Meads, Epigram was invited to have a sneak peak inside the University’s most adventurous project to date.

EpigramDana Andrea Datu

The report also highlighted some evidence of universities developing their own AI tools for this process.

According to Professor Lawrie Phipps, co-author of the report, there is ‘an urgent need for strong governance and shared standards.’ Currently, there is a ‘growing inequity between those able to develop bespoke systems and those reliant on public tools.’

The University of Bristol is one of the HEIs that has taken recent steps towards developing its expertise in AI.

In July, the UK's fastest supercomputer, Isambard-AI, was launched at the university. On November 21 the university announced that it is set to develop a National AI data facility, which will function as a digital library for researchers across the UK. This will be located next to, and work alongside, Isambard-AI.

University of Bristol to develop new ‘digital library’ as part of UK’s AI revolution

The plans come after the launch of Isambard-AI at University of Bristol in July, and supports the wider strategy of the University, and the West of England, to develop their position as a hub for AI innovation.

EpigramAimee Anderson

The report produced a survey of 400 academics and professional service staff opinions on AI being used within REF2029. 54-75% of respondents were ‘strongly opposed’ to the usage of AI within REF.

The report also interviewed university Pro Vice-Chancellors. Feedback included concerns that we are ‘in a bit of an AI bubble at the moment’ and that ‘we don't really trust [AI ] enough in order to really start backing away from conventional tools.’

Some Pro Vice-Chancellors were more supportive of using AI within REF. One respondent said: ‘this is the future... we need to lean into it.’

Within the report, one HEI REF Lead respondent said that ‘the best way to keep your job and not get replaced by AI is to be the person that can manage AI.’ 23% of academic and service staff survey participants supported the use of AI by universities for case studies that assess the impact of their research.

The report urges 'the development of formal governance frameworks, sector-wide standards for transparency and disclosure, and equitable access to safe, secure AI tools.'

The report recommends that all universities establish and publish a policy on the use of AI for research purposes. It also indicates that relevant staff should receive full training on the use of AI tools.

The report urges ‘the development of formal governance frameworks, sector-wide standards for transparency and disclosure, and equitable access to safe, secure AI tools.’ This is necessary to preserve ‘the integrity of UK research.’

Featured image: Epigram / Hanno Sie