The adoption of artificial intelligence across higher education requires boards to have an understanding of how the technology is being used now, with an eye to future developments, and the risks and opportunities it presents.
Within governance itself, AI input into decision making and strategic thinking is increasing, raising questions about visibility, effectiveness and good judgement.
Recent developments in sector guidance have reinforced this shift. The new Committee of University Chairs (CUC) Higher Education Code of Governance (2026) places strong emphasis on adaptability, accountability and impact. The Advance HE Big Conversation: Shaping the Future of HE Governance has underscored the growing expectation that governance must become more agile, more inclusive, and more connected to real-world complexity.
New Advance HE, Association of Higher Education Professionals and Edge Hill University guidance, Using AI to Support Governance Effectiveness, has been created with this evolving context in mind. AI systems can influence decisions, shape outcomes, and introduce new forms of risk. Traditional governance approaches, if applied passively, risk being outpaced by the speed and complexity of these developments. Instead, the guidance says, governors must demonstrate “strong professional curiosity, ethical awareness, and practical engagement”.
The guidance highlights key areas to inform and guide future practice, including connecting AI to real decisions, strong data governance and applying performance assessments and evaluations.
A commitment to ongoing professional development is a starting point, the guidance suggests. Increasingly boards are being exposed to activities aimed at keeping governors up to date with AI developments and how it can and is being used in the sector.
A governor in Scotland cited a recent court away day where a guest speaker was invited to talk about AI: both its wider impact and its application to HE. A series of court information sessions has also included discussion of what the university is doing “in the AI space”.
Meanwhile, AI’s integration and impact on student services and teaching and learning is a major focus for the executive generally, led by senior managers who have responsibility for it.
“Institutionally we have done a lot of work in professional services on AI and automation and how that affects the way we work,” said the staff governor. “I’m on a steering group overseeing that work and its progress is presented back to court in regular papers.”
A governor at a for-profit institution which is investing substantially in AI said there have been “a lot of AI discussions at board level” around AI.
According to this governor, the challenge is that “there is a group that understand pedagogy but not the technology, and a group that understand the technology but don’t know much about teaching.”
He added: “We need to bring those two things together and at that intersection, it is a collective exercise. There are vague ideas of what the university of the future will look like, but it is a learning process.”
The board member runs monthly online meetups that are attended by the executive but also by some board members, where presentations on AI have been given, from both internal and external speakers.
“I think we are probably doing as much as anyone and have as good a handle as you can have, but you have to be realistic,” said the governor. “I spend a lot of my time hanging out with ‘techy sorts’ and they don’t know exactly where the tech is going or how it should be used in education. It is an intrinsically uncertain area and we are all on a voyage of discovery.”
Meanwhile governors at a Russell Group university, which has made recent investment in its digital infrastructure, including AI capability, are due to have training on using integrated AI in September. A governor at this university has already undertaken his own research and training outside of his university role.
From recruitment chatbots to using AI in assessment and predictive analytics and forecasting, governors are getting to grips with how AI tools are being utilised internally, considering the implications, and formalising how this is monitored.
“There is a lot of visibility about how it is being used in the institution,” said one governor. “There are week to week updates on this stuff. It is important to reflect on what AI is doing, whether that be chatbots or other types of machine learning.”
At his institution, predicative analytics look at student engagement and help the support team target interventions. AI uses historic data about non-continuation in a bid to spot the early signals in current cohorts to stop it happening.
“We have been doing it for a few years and now have a whole team on predicative analytics,” he said. “We talk about it at board level, but the challenge is really understanding what it means if someone is given a 70 per cent chance of dropping out. Is it accurate? What data have they used? Is it fair if you are using it to target intervention? These are all important questions.”
There is “good visibility and healthy debate”, says the governor, but he not sure that it is backed by a full appreciation of “how it works and what it means”.
Governors also need to be aware of other potential uses that might be less positive.
“If you were to use that information to manage admissions, for example, there is a good bet that you would be negatively affecting people’s life chances. We are not doing this of course, but there are a whole lot of ethics around its use that governors should be aware of.”
Visibility and understanding are vital if governors are to scrutinise ongoing performance and evaluate the efficacy of AI in decision making. The new Advance HE guidance on AI and governance says continuous oversight is often built on “simple, repeatable behaviours”, such as regularly asking “is this still working as expected?”, rather than complex structures.
This approach is demonstrated by one governor at a Welsh university who said that when presented with financial modelling, for instance, he always asks if it is AI calculated.
“My own background means I will always ask how a model is built,” he says. “Even if I am indifferent as to whether it is AI or an older-fashioned model, I’m interested in how the engine works.”
Part of that scrutiny of AI application means being assured that it is improving outcomes and/or supporting governors and the executive to meet statutory duties and regulatory obligations. For instance, is the proposed AI-use linked to clearly defined business outcomes or strategic aims, is there an AI-use pros and cons mechanism, and are there reviews of AI-use?
One institution researching the use of AI in marking and assessment is a good example. It has potential strategic benefits in areas such as staffing, workload and student numbers, as well contributing to improvements in marking accuracy and timely feedback.
“At the moment the research is showing that where AI is marking, there is a high level of agreement between machine and human,” said a governor. “That is positive, not least because all academics hate marking. We will be looking at how much time can be saved if AI is used for it instead. We will also be looking at how that time should be reinvested, for instance more face time with students or the capacity to take on more students.”
Once in place, the marking policy and the benefits it produces would be subject to ongoing board scrutiny, he said.
As in wider society, there are ethical considerations if AI is employed in universities to undertake work that is currently carried out by staff.
A governor in Scotland said the board was “very conscious” that some colleagues are “anti-AI”. More generally, she believes there is professional scepticism around its use at board level, and that this is as it should be.
“This time last year everyone was interested in the opportunities that might be provided by AI; how might this help given difficult finances. But I think I am hearing a lot more conversations about the costs of AI sustainability in our strategy, for instance,” she said.
According to one governor, it is “really important that sceptics get involved”.
“We have to train our students in AI because we can’t send them out into the world without that, but if we only listen to the evangelists, we will miss the pitfalls,” he said. “Sceptic needs to be involved to temper the overenthusiasm.”
Professional scepticism should be built into the audit processes, according to a Russell Group governor.
“When it produces an answer, we need to ask it, ‘what would your answer be if I’d said this?’ or ‘what would the other answer be to this question?”, he said. “In practical terms, you have to build in an audit system; you need to put something in the system that can test the system.”
Concerns about AI range from its tendency to “hallucinate” and produce inaccuracies, the impact on net zero targets of its energy and water consumption, and cybersecurity threats, which it can both protect against and make institutions more vulnerable to.
While risks are present, they should not be overblown, according to a board member at a for-profit provider. He argues that while energy use was a genuine concern, there is a certain amount of misinformation on this.
All governors who spoke to Advance HE said that AI had been added to their university risk register and that audit and risk committees report regularly to full board meetings on the item.
Risk assessments covered external factors, such as the risk AI might pose to graduate jobs, as well as internal risks – from staffing, competencies, recruitment, cybersecurity, and, as one governor puts it, “whether it will mean nobody wants to go to university anymore”.
A Welsh governor said that AI was “absolutely part of our risk register” in a bid to guard against “perverse outcomes”.
“If the AI algorithms let in so many of a certain profile of student to study ancient hieroglyphics, for instance, it might fill the course quite perversely,” he said, “which paradoxically makes human oversight even more important. The risk is always there that you may have set the AI the wrong questions without knowing, or you may have set it to take account of certain criteria, but it is ignoring other criteria that turn out to be equally important to the bigger picture.”
Overarching institution-wide AI policies were also seen as important to provide a consistent framework for how the technology is used across teaching and learning, research, administration, and student support, to cut down on risks, confusion and conflicting practices.
A Scottish provider has established a group that is developing its AI policy to be “reflective of what we know now” that is due to go through the various governance structures.
Another board member said setting out expectations across the university was essential, but also made the point that policies needed to be reviewed regularly “in light of fast-moving developments”. “We have a statement of use of AI setting out expectations not only for students but for staff about how they will and won’t use it and the level of transparency of use,” he said. “It is about good and bad uses of AI and discovering which is which. It will continue to be refined as we learn more and as the technology becomes even more sophisticated.”
