Introduction
There is no shortage of discussion about the effect of artificial intelligence on employment. Some predictions are undoubtedly exaggerated, but the underlying question is legitimate. If AI enables organisations to undertake the same amount of work with fewer people, it would be surprising if there were no consequences for employment.
Higher education is an interesting case because we talk a lot about AI changing what universities do. We debate its use by students, its implications for assessment, its effect on research, and how academics should incorporate it into their teaching. We talk rather less about another possibility. Could AI mean that universities eventually need fewer employees?
Universities are already losing jobs
We need to be careful before answering that question because higher education is already experiencing substantial workforce reductions.
In several countries, universities are under severe financial pressure. Falling or changing student demand, dependence on international recruitment, rising costs, constrained public funding and demographic change are all contributors. In the UK, for example, large numbers of universities have introduced recruitment freezes, voluntary redundancy schemes or other measures to reduce staffing costs.
These are serious issues that are affecting many people, and the focus on this aspect of higher education should not be minimised, or confused, by the effects of AI. We should not suggest that AI is responsible for every university redundancy.
However, if a university introduces AI at the same time as it reduces its workforce for financial reasons, how would we know whether AI had played any part in that reduction?
Nobody needs to say that AI took their job
Imagine that a professional services unit in a university employs 20 people and is told that it needs to reduce its costs.
Five years ago, reducing the team to 17 might have meant accepting that less work could be done or that service levels would fall. Today, the calculation may be different. AI might help the remaining 17 employees handle enquiries, analyse information, draft documents, produce reports and undertake other routine activities more efficiently.
The official reason for losing three posts might still be financial restructuring. But AI may have made the restructuring possible.
There is an even less visible mechanism. Suppose somebody leaves a university and their manager would previously have recruited a replacement. Instead, the university asks whether the work can be redistributed among the remaining team, supported by AI and automation.
Nobody has been made redundant. There is no announcement that a job has been replaced by AI. There might not even be a conscious institutional decision to reduce the workforce because of AI. The job simply disappears.
If this happens thousands of times across universities, conventional measures of “AI job losses” could considerably underestimate what is happening.
Where might the impact be felt first?
My suspicion is that professional services roles may initially be more exposed than mainstream academic roles, although there will be considerable variation within both groups.
Consider some of the work undertaken across universities: answering routine student enquiries, processing applications, preparing reports, producing marketing materials, analysing data, drafting correspondence, translating documents, recording meetings, supporting research administration, responding to basic IT queries and completing routine financial or HR processes.
Many of these activities can already be partly automated or substantially accelerated using AI. That does not mean that the people undertaking them are unnecessary. Most jobs consist of multiple tasks, some of which are much easier to automate than others.
But the economics potentially change when AI allows five people to undertake work that previously required six.
EDUCAUSE’s 2026 research on AI and work in higher education found that more than 70% of surveyed higher education professionals were already using AI tools daily or weekly. Most institutions represented in the survey had some form of work-related AI strategy, and many were focusing on upskilling or reskilling their existing workforce. AI is therefore no longer an experimental technology sitting on the margins of university administration. It is becoming part of normal work.
The employment consequences of that change remain much less clear.
What about academics?
Academic roles may appear more protected.
AI can help academics develop teaching materials, summarise literature, analyse data, draft correspondence, produce formative questions and undertake many administrative activities. It can potentially reduce the time required for some aspects of assessment and feedback.
Yet an academic is not simply a collection of these tasks. Teaching involves relationships, judgement and disciplinary expertise. Research involves deciding which questions matter, interpreting evidence and generating new ideas. Supervision, mentoring, academic leadership and collegiality are similarly difficult to reduce to an automated process.
Nevertheless, it would be premature to conclude that academic employment is immune.
The American Association of University Professors has already identified possible job loss, deskilling, work intensification and greater reliance on contingent academic appointments among the risks associated with AI. Its concerns do not demonstrate that widespread AI-driven academic job losses are occurring, but they show that the possibility is being taken seriously.
There might also be an indirect effect. If AI allows an academic to teach more students, supervise more projects or perform administrative work more quickly, a university under financial pressure may eventually ask whether it needs the same number of academics.
Again, the resulting decision might be presented as financial restructuring rather than AI substitution.
Are knowledge workers protected?
There is another assumption worth questioning. Universities employ highly educated people doing knowledge-intensive work. We might therefore assume that our jobs are safer than those involving more routine activities.
Recent evidence suggests we should be cautious.
Research published by Malaysia’s Institute of Strategic and International Studies in June 2026 examined loss-of-employment data and occupational exposure to AI. Among its findings was that tertiary-educated employees working in AI-exposed occupations experienced substantially higher rates of employment loss. Early-career workers were also particularly exposed.
That study was not specifically about universities, and we should not pretend otherwise. But its broader message is important.
AI does not necessarily threaten only routine manual occupations. Generative AI is unusually relevant precisely because it can perform parts of the cognitive and information-processing work undertaken by educated professionals.
Universities employ a great many such professionals.
Perhaps we are asking the wrong question
At present, I can find little convincing evidence that universities are making large numbers of people redundant specifically because AI has replaced them.
That is important. We should not manufacture a crisis for which the evidence does not yet exist.
But neither should we conclude that higher education is somehow immune.
Perhaps instead of asking, “How many university employees have been made redundant because of AI?”, we should be asking different questions.
Are vacancies being left unfilled because AI enables existing employees to absorb the work? Are administrative teams becoming smaller? Are universities replacing fewer people who retire or resign? Are new roles being created at the same rate as before? When universities restructure because of financial pressures, does AI give leaders confidence that fewer people can deliver the same services?
Those effects would be much harder to see.
Universities have spent the past three years asking how AI will change our students, our teaching and our research.
Perhaps it is time we also asked what it will do to us.
I publish two free newsletters each week
Governance, Leadership & Research
A weekly brief for higher education leaders and researchers examining governance, leadership and research in practice.
https://buff.ly/GhsZqVl
Publishing with Integrity
Examining scholarly publishing through an ethical lens, challenging assumptions and supporting academic career development.
https://buff.ly/WiG1yFb
If you subscribe, you will be alerted whenever I publish a new edition.
About the author
Professor Graham Kendall is Vice-Chancellor of GlobalNxt University, Malaysia, and an Emeritus Professor at the University of Nottingham. He has held senior university leadership roles for more than 15 years, including responsibility for research, knowledge exchange and university-wide operations. He has published more than 300 peer-reviewed papers and writes regularly about higher education leadership, research, artificial intelligence and scholarly publishing.