Gone are the days of leafing through dusty journals in search of obscure references. Today’s students have turned to generative AI platforms for help, using AI-powered tools to compile resources all the way through to crafting entire assignments. But despite strong uptake from students, universities have been more cautious. According to new research from PA Consulting, only 10 percent of university vice-chancellors feel they are currently doing enough in this area.
In a sector beset by funding constraints and uncertain demand, AI could help to achieve the reform that 94 percent of vice-chancellors say are needed for the sector’s survival. But the majority (78 percent) of vice-chancellors are still in the planning stages. Now that over nine in ten of UK university students are using AI tools to support their studies, there’s an onus – and expectation – on universities to incorporate AI into operations at all levels.
Universities are starting to dip their toe in the water – Gen AI is being used to support curriculum design, rubrics, and assessment plans, as well as enabling formative assessment at scale. Researchers are using Gen AI to help create funding proposals, assess funding opportunities, and support research activity itself. In professional services, AI can aid query management and increase workflow automation through improved data enablement. However, universities aren’t realising the full potential. What do they need to do to make the shift?
Prioritise the right AI use cases
Universities want to move past the “AI is the answer, what’s the question?” approach they experience from suppliers towards a more value-focused, nuanced approach – one that allows them to align their AI roadmap with their strategy.
This will allow them to avoid the scattergun effect, with isolated pilots taking place under the guidance of singular AI pioneers. Without an understanding of why, where, and how AI can help, these pilots often wither and die, burying potentially promising AI use cases with them. A structured approach is needed, with clear criteria relating to specific tools, use cases, and scalability plans.
Strategies also need to consider the potential impact of AI on learning and teaching – not just the content, but also the fundamental implications on knowledge-based learning cultures. Different use cases will be better suited to different universities, so aligning use case development with particular strategic outcomes is crucial. This will deliver tangible, strategic benefits early on, providing compelling evidence for greater AI adoption across the organisation while inspiring the wider sector.
Finding the right use cases starts with mapping out potential applications. For example, the University of Warwick recently hosted an AI Innovation Challenge to prioritise potential AI use cases. Four AI agents were developed for four priority cases, which were demonstrated to a panel of senior stakeholders from the University. One agent focused on scholarship scoring moderation, reviewing applications against defined criteria to shortlist candidates and save significant time per application round. The challenge-based approach uncovered opportunities while laying the foundations for structured development.
Think ‘AI first’ – but keep humans in the loop
AI can deliver real improvements across higher education institutions, aiding recruitment, admissions, essay marking, and policy amendments. But deriving true value from AI adoption takes a transformational mindset. Many institutions are still stuck in the ‘buying a faster horse’ phase, which automates existing processes, taking advantage of faster processing to speed up previously human-led processes. But this won’t address the issues that are likely to cause AI gains to plateau, such as siloed data, inefficient processes, and a delivery model that doesn’t compliment wider strategic goals. It also risks over-automation – it’s key to retain human-in-the-loop decision-making, which protects ingenuity and nuance.
To achieve real efficiency and experience gains, universities need to redesign how they deliver specific services through an AI lens. In practice, this means addressing disparate data, linking up university systems, and setting the right safeguards and security controls. It also means working out where people will need to step in – either to correct an algorithm’s decision or to provide all-important human support.
An AI first mindset can be supported by external partnerships. Microsoft has made Copilot available to all higher education students and faculty as part of Microsoft 365, while Google has added Gemini to Google Workspace for Education. Tapping into these tools is a no-brainer, while specific AI platforms such as Remesh can help universities to uncover student perceptions of diversity, inclusion, and wellbeing through anonymous feedback, removing the fear of judgement or repercussions to discover previously unexplored issues.
Balance competing priorities
Universities are under ever-greater pressure to deliver efficiently to manage expenditure and protect resources. At the same time, they are expected to be more sustainable and ethical in how they deliver services to student cohorts and various educational partners.
While AI can reduce administrative burdens and save considerable time and effort across operations, learning, and research, it comes with a sustainability cost – particularly water and electricity. Universities need to be able to evidence that using AI won’t compromise sustainability obligations, and that sustainability has been considered as part of their AI adoption strategy. To do this, they can develop a prioritisation matrix that brings different priorities together, balancing them across the different areas with overarching governance mechanisms.
Don’t be afraid of AI
If used sporadically and without a governing strategy, AI transformation will inevitably fail – especially in resource-constrained environments. However, with the right strategic focus, alongside the use of internal and external experts in relevant roles, AI offers an unmissable opportunity to achieve the reforms that are fundamental to the sector’s survival. The willingness to take advantage of AI is there – universities need to dive in and make it real.
