GenAI and the University’s Favourite Magic Trick
Generative AI has clearly arrived in universities in the traditional manner: on a wave of hype, inside a PowerPoint, and accompanied by several people who say “transformative” as if it were both a plan and a moral virtue.
GenAI is supposed to personalise learning, support students, reduce workload, enhance productivity, improve feedback, widen access, deepen inquiry, and possibly make a decent flat white. Any minute now it may also begin repairing the plumbing and healing the spiritual wounds caused by strategic planning.
Now, some of this can be true. GenAI can be useful. It can help draft, summarise, rephrase, brainstorm, translate, and simulate. It can be genuinely helpful to students and teachers. I’m less confident about healing spiritual wounds, but that is not the issue. The issue is that GenAI is not arriving in a healthy institution. It is arriving in the modern university and the modern university has a special talent.
Its great achievement is no longer the production of knowledge, though it still occasionally does that between meetings. Its true genius is the creation of managerial headroom: the ability of senior leaders to free up time for strategic thinking by ensuring that nobody lower down has any time left to think at all [1].
This is a superb administrative innovation. It works by increasing workloads, enlarging classes, casualising teaching, multiplying compliance rituals, and then announcing that the institution is now more agile. One group gets “capacity for strategic oversight.” The other gets burnout, browser tabs, and a reminder about mindfulness.
As Hannah Forsyth argues, managers gain headroom by hollowing out the labour that gives the university its legitimacy in the first place. Teaching, research, intellectual life, pastoral care, and actual thought. These are the things the institution points to with enormous pride while systematically treating them as expensive obstacles to efficient administration. The university brochure says: we change lives. The internal operating model says: please log that in the portal by 4 pm Thursday.
This is where GenAI becomes really interesting, in the same way a new species of mould becomes interesting when introduced into an already damp basement.
Because GenAI might be used to enrich education. It might help teachers explain difficult ideas in different ways, help students rehearse concepts, support drafting, generate examples, and create forms of intellectual play that were previously too labour-intensive to attempt. But it might also be used for the thing universities are now best at: converting a promising tool into a mechanism for buying flexibility at the top by thinning life at the bottom. That is the concern.
GenAI can be sold as support while functioning as substitution. It can be marketed as liberation while operating as standardisation. It can be introduced as a way to “free up academics to focus on what matters most,” which is one of those phrases so perfectly engineered to avoid meaning anything that it should be preserved in a museum.
What often happens in practice is simpler. Feedback becomes templated. Student support becomes triaged. Teaching becomes content delivery with a chatbot garnish. Judgment becomes analytics. The odd, slow, human business of education gets translated into things that can be counted, monitored, accelerated, and reported upwards [2]. And that is not an accident. It has a history.
Michael Pusey (1991, 2018) among others, called it economic rationalism. That long and not only Australian habit of taking public institutions, distrusting the people who actually do the work, and then surrounding them with managerial systems until the systems become the work. The language is always efficiency, accountability, performance, modernisation. The result is usually more forms, fewer humans, and a strange institutional confidence that anything which can be measured must therefore be important [2].
Economic rationalism did not just change how institutions were run. It changed what they thought intelligence looked like. Professional judgment? Messy. Relational. Slow. Hard to graph. Suspicious already. Managerial calculation? Splendid. Neat. Transportable. Looks marvellous in a dashboard. Makes everyone feel that something is being done, which is the highest form of action in many large organisations. GenAI slots into this world rather neatly, way too neatly.
If audit culture made academic work legible to management through metrics and reporting, GenAI may make the work itself more available for formatting, monitoring, slicing, and substitution. More teaching can be standardised. More interaction can be automated. More labour can be made visible in ways that are useful chiefly to people who have not done it for years. This will all be described as agile innovation.
It will not, naturally, be described as what it may sometimes be: a way of extracting yet more managerial headroom from an institution already running on depleted goodwill, casual labour, and the final fumes of professional pride.
That is the central absurdity. Universities still derive their prestige from scholarship, teaching, and inquiry. They still talk as if they are communities of thought. They still bask in the reflected glory of serious intellectual work. But more and more they organise themselves as if thought were an unfortunate side effect of a larger administrative enterprise [3].
They boast about the orchestra while gradually replacing the musicians with workflow software. And then, when the music becomes thin and strange and somehow entirely in E minor, they commission a review into sonic resilience.
To be fair, not all management is fake, not all strategy is nonsense, and not all uses of GenAI are corrosive. Some are sensible. Some are overdue. Some are even useful and sometimes good. But that is not a reason to swallow the sales pitch whole like an enthusiastic pelican. The question is not whether GenAI is useful. The question is: useful for what kind of university?
If it gives teachers and students more room for thought, feedback, experimentation, and care, excellent [4]. That is worth having.If it gives senior management more room to manoeuvre while making teaching thinner, support more impersonal, and academic work more standardised, then we are not looking at educational renewal. We are looking at economic rationalism with a forgetful chatbot. Which, like many technological upgrades, is essentially the old problem wearing a more colourful blazer and using the word “ecosystem” far too often.
Notes
1 Forsyth, H. (2026, April 19) Uni managers gain 'headroom' by fucking everything up. https://hannahforsyth.substack.com/p/uni-managers-gain-headroom-by-fucking
2 Bigum, C. (2026, February 27). So Long, and Thanks for All the Metrics. https://chrisbigum.blogspot.com/2026/02/bibs-bobs-34.html
3 There is an opportunity here for someone to write Seeing like an Australian University and borrow from a 2025 Hollis Robbins post that details Seeing Like a State University (in the US).
4 Assuming that the epistemically uninsured are not running things.
References
Pusey, M. (2018). Economic rationalism in Canberra 25 years on? Journal of Sociology, 54(1), 12-17. https://doi.org/10.1177/1440783318759086
Pusey, M. (1991). Economic Rationalism in Canberra. A Nation Building State Changes Its Mind. Cambridge University Press.
