Introduction
Higher Education in the UK is in the throes of a three-fold existential crisis.
The first faultline concerns the much-discussed chronic crisis of university profitability under the government-backed funding system. In 2011, a home undergraduate student fee of £9,000 obtained a surplus of around £2,000 per student against the average cost of tuition.
When the cap on student recruitment was abolished in 2014, English universities jumped headlong into a free-for-all competition to expand, in a war of ‘all against all’. They borrowed heavily, pouring millions into long term capital projects. In addition to new buildings and even campuses, they expanded speculatively, including investing in London and overseas campuses. Scottish, Welsh and Northern Irish universities followed suit more tentatively. But by COVID lockdown in 2020, UK universities had borrowed in the billions. Recent interest rate hikes have had a differential impact across HE institutions, but even the savviest VC signed up for loans predicated on an unrealistic model of continual long-term growth in student recruitment and fee surpluses.
The architects of this HE market, the Conservative Government and its successors, refused to increase home undergraduate fees with inflation. University employers responded by a number of measures: expanding student accommodation, recruiting overseas students and growing ‘unregulated’ fee sectors such as taught Masters — and, importantly, holding down the staff pay bill.
As a result, the proportion of university budgets spent on staff fell to a record low of just above 50% by 2023. Yet despite cutting staff pay by nearly one third in real terms over the last decade and half, English universities today claim that they are teaching home students at break-even or a loss.
Even when staff costs fall dramatically, Vice Chancellors continue to see staff as a target for cuts. From 2021 to 2023, the UK was hit by a period of very high inflation. By August 2023, staff pay had been cut by 11.7% against RPI (7.8% against CPI) in just two years. Pre-92 universities reaped an additional unearned windfall in early 2023 of another 4% of the wages bill, after the UCU trade union successfully defended the USS pension scheme.
Yet these windfalls have not prevented VCs across the UK seeking wholesale job losses to restructure their offers to students. In 2023/24, many employers announced a call for 10% cuts in staff. And the top of the 2024/25 hit list is a roll-call of pre-92 universities, including Dundee, Newcastle, Brunel, Cardiff, Sheffield, Durham, Nottingham. A pay offer of 3% below inflation in 2025/26 is unlikely to stop redundancies.
Universities in Scotland, Wales and Northern Ireland, who never received these higher fees, had reacted to the market in England from the get-go by expanding their international student recruitment operations. It was not long before English universities followed suit.
Unsurprisingly, international student recruitment has also proved to be an impermanent solution to UK university funding. Leaving aside the interaction with the second aspect of the present crisis, which we will come to, where anti-migrant rhetoric from governments leads to restrictions on overseas student numbers, countries such as China are developing comparable universities of their own.
The Chinese student market situation is a case in point. Having been a long-standing source of high-fee overseas students to the UK, with some 90,000 in 2014/15 growing to 154,000 in 2022/23, there are signs that numbers may be falling. Some 100,000 students were offered visas for 2025-26 study by June 2025. A slowdown in the Chinese economy may also mean fewer Chinese middle class parents prepared to invest in a Western education for their offspring.
The global repositioning of the US under Donald Trump towards seeing China, rather than Russia, as its main competitor, is having uncertain consequences for the UK. One might speculate that had Keir Starmer steered sharply away from anti-migrant rhetoric and distanced himself from Trump, it is possible that international student numbers in the UK would have risen sharply as they did post 9/11, with the UK rather than the US becoming the destination of choice for English-language student instruction. But it was not to be. Starmer preferred to cosy up to Trump and lurch sharply to the right, squandering a major opportunity. Indeed, Starmer and his government has treated the Vice Chancellors with little more than Thatcherite contempt, telling them to ‘get their own house in order’ to balance budgets.
Trump and the UK far-right
This leads me to the second faultline, namely the rise of so-called ‘populist’ far-right parties which are instinctively anti-science, anti-intellectual and anti-universities. Trump’s attacks on Harvard and Columbia are a piece with his purge of the Center of Disease Control, Voice of America, etc., proving the old adage of the indivisibility of freedoms. Despite some notable protests, US university managements have largely capitulated to Trump’s witch-hunt. The defence of academic freedom for universities and staff is inextricably bound up with a necessary defence of science, public information and an educated critical society. Independent critical science needs independent critical media.
The UK is not immune from these trends. The writing was on the wall when the Conservative Party rallied against “woke science” (sic) during its 2023 Party Conference. When the Grand Old Party of the British Establishment decides to attack the Enlightenment to curry political favour, universities are obliged to realise that they had become fair game in politicians trading in gutter politics and abuse.
The historic truce between universities and state, parameterised by institutional autonomy and academic freedom, is under threat.
Nigel Farage’s ‘Reform’ project is likely to emulate Donald Trump by attacking universities for teaching anti-racism or challenging transphobia, but this will be just the start. At the very least, we can expect open anti-science propaganda and political interference in universities, the health service and the media. How far this succeeds will depend on the breadth of resistance.
Trump’s cards are already on the table. Farage’s are unlikely to be different.
Artificial Unintelligence?
The third faultline of the present crisis is ‘AI’. Or to be more precise, the corporate investment in and development of Large Language Models (LLMs) and other inductive algorithms, coupled to a huge investment in large data centres since 2022. There are some signs the first wave of trillion dollar investment in the US, amid competition with China, is creating an investment bubble that will burst.
Like the Dot Com bubble, the rapid expansion of capacity without profitable use cases is liable to trigger panic from investors, and a temporary retreat from the boosterism of the last year. But like internet bandwidth, AI and similar technology is here to stay.
‘AI’ is having two impacts on higher education worldwide. The first, a crisis of assessment, has been widely commented on within the academy. Should lecturers embrace AI in their teaching, and how might assessments be ‘AI-proof’? Important though these questions are, a much bigger issue is looming over education.
This second dimension is beginning to be felt as a drop-off in the market for graduate jobs, especially, it appears, in so-called ‘AI-adjacent’ subjects. Whereas this may be a result of simple timing (a downturn following a recent hiring spree) the broader risk is that graduates are leaving higher education ill-equipped for the potential AI replacement of skills and knowledge they have acquired the traditional way.
A recent study of the US job market indicates that ‘experience’ is being prioritised by employers over courses perceived to be concerned with mere ‘acquisition of facts’. Meanwhile, one of Google’s AI leaders warns that the PhD study of law and medicine may have limited market value.
Nonetheless, Higher Education cannot — indeed must not — be reduced to graduate job training. When we say that independent science needs independent media, we imply the public must be sufficiently educated to engage with it.
A scientific correction to AI boosterism starts from the recognition that ‘AI’ is not a substitute for actual knowledge. Machine learning interpolates lines from points and guesses gaps in data. It is limited both by the quality of the original data and by the framework it applies. Where data is dense and underanalysed, AI can speed up analysis and make real gains, such as discovering new drugs. But AI cannot make up for truly sparse data. And even when new drugs are discovered, expertise and testing is required to validate them.
This means two things: expert knowledge will become more important, not less, in an AI-augmented world. And more research will be needed to sample missing data.
When Chat-GPT first became popular, ‘hallucinations’ were obvious. But as AI improves, a greater level of expertise will be required.
Indeed, ‘AI’ will certainly be used as a form of product promotion and advertising. Google’s monopoly position as search engine of choice is already being bypassed by AI services. But AI-driven product promotion, insinuated into an individually-created response to a query, is likely to be more difficult to identify, regulate and correct. As Geoffrey Hinton recently commented, AI is already much more persuasive than humans. Investors in the AI bubble are expecting returns.
Addressing the Triple Threat
These three faultlines are coalescing into a perfect storm, a ‘triple threat’ to UK HE. What I mean is not that they represent three independent co-occurring threats, which they undoubtedly do, but rather that these processes are interacting in particular ways, such as when the far right use pseudo-scientific reasoning and AI algorithms to promote their agendas, from vaccine denial to Donald Trump’s infamous Gaza video, or in algorithms for social media promotion. Similarly, the UK Higher Education sector is facing the challenge of AI in ways that, say, the state-backed German academy, without the UK sector’s student fee-paying market-dependence, is not.
We urgently need to create the conditions for an open debate on what HE is for, in a world where AI tools are a fact of life. This debate needs to begin within the academy: if we don’t take on this task, others will claim to speak for us.
As a general principle, we need HE for humans to control AI. As we note above, inductive reasoning is faulty, at best exploratory. The importance of Higher Education in equipping the next generation to grapple with AI failures, misrepresentations and hallucinations is greater than ever.
Longer term impacts are difficult to predict, but the impact on HE teaching, as on wider society, is likely to be massive. A wholesale curriculum review is inevitable.
But a key democratic question is who drives this review? It seems to me that academics and professional bodies are best-placed to do so, and have the necessary skills and capacities to do so.
It would be a disaster to leave this to ‘the market’. Vice Chancellors and senior managers seeking mass redundancies as they try to reposition universities in a rapidly changing marketplace have no real idea of what is happening to the sector, no strategic thinking and no moral authority to assert the case for Higher Education to the wider public. Employers and politicians are similarly unaware.
Such an exercise will require us reasserting academic control over course design from the managers who spent the last fifteen years re-imagining courses to sell on the HE market and slashing jobs. Chasing the market is what got UK HE into its current mess!
First steps, like adding ‘AI’ modules into existing courses, are a start, but we can expect AI to change professional disciplines profoundly. Tomorrow’s graduates will need ‘traditional’ factual subject knowledge, practical skills, but they must develop a critical understanding of the potential and limits of generative AI.
For a start, I think that we will all need to ‘up our game’ and teach critical thinking in both a general and subject-specific way.
The political bad-actor scenario also reminds us that societies can go into reverse. Only academics speaking up for their disciplines and their universities, allied with an educated democratic public, can deal with hostile parties and governments led by the Trumps and Farages of this world. In defending our disciplines, we will also need to persuade the wider public of the necessity of such a defence. But this argument can be won, as public attitude surveys such as HEPI 2022 continually remind us.
The Triple Threat brings one final element into sharp focus.
How will such a debate be organised in practice?
Some points seem obvious.
- The debate can be begun by academic staff demanding their trade union branches host open meetings. At the moment, staff are aware of the crisis but individuals cannot create an organised response. That requires a trade union.
- This debate can help create the space for more detailed discussion about curriculum. Subject-level reviews must be conducted at course leadership, professional bodies and university boards, Senates or Academic Boards.
For the academic unions, UCU and EIS, the question becomes straightforward. The financial crisis imposed on UK Higher Education can only be addressed by a shift away from dependence on student tuition fees towards a sustainable funding model involving increased direct state support. This will require both an industrial and political campaign.
Central to such a campaign involves a restatement of the case for Education as a public good. Higher Education is necessary for a democratic society in which AI exists. The alternative is the subordination of humanity to AI and the corporations that wield it.
As the NUS used to say, ‘if you think education is expensive, try ignorance.’
Sean Wallis, University College London
