On 13 March 2026, the Geelong Advertiser reported outrage among students as they realised they were paying full fees to Deakin University for lectures delivered by AI. One student’s comment captured the sentiment: ‘[Universities] need to cop a jaw-shattering punch from the regulators… [t]heir laziness and greed has become even worse in recent years’. But where are the regulators, who is accountable for the greed, and will the bot hires stop, or are they just getting started?
A month earlier, some academic friends sent me a link to Einstein.ai, a Large Language Model (LLM)-powered tool that purported to be able to log into the common university digital platform Canvas to complete and submit all their work for them. One of the academics who sent me the link quipped: ‘game over’. It was indicative that, according to some reports, faculty members themselves have been the most eager users logging into this tool.
Although the site publicising the US-based Einstein.ai has been shut down for now, one feels we are approaching the sad finale of this model of higher education. Across Australian universities, local leadership is showing itself very willing to meet the demand for automated agents whose essential purpose is to avoid hiring and respecting those who think and teach, because this is more ‘efficient’. In 2026, the university is a content-delivering credential factory; if credentials can be delivered more efficiently by a bot, then so be it.
In the new AI-powered meeting of universities and students, scholarship and pedagogy are collateral damage: wisdom appears to be a dead pursuit, especially if it’s embodied in the slow and idiosyncratic ‘meatware’ of a person. Knowledge is for chumps, especially if you have to show up and grind for it. Operating like this might be a winning game for some, especially for the consultancy-aligned administrators spruiking the efficiency dividends on their LinkedIn profiles, and for students acting in bad faith, who pass without learning anything. When a university does something as cynical, lazy and greedy as wheel out a bot to teach you, at full price, perhaps the sensible thing to do is to source and deploy a counter-bot of your own, carry on with your life and not think about the fifty grand your new piece of paper just cost you.
This systemic state of play predates AI by at least a dozen years, but AI is making things much worse, very quickly. This is because AIs win at the universities’ marketed game of ‘work smarter, not harder’, ‘easily fit your studies around your life’, and ‘never sacrifice quality for convenience’. It’s not hard to see from a student’s point of view that using AI is a prudent and rational response to the hollowing out of the universities. We have arrived at the 24/7, flexible online delivery of the qualification without qualities.
In understanding how we got here, much has been written, extending back to the Dawkins reforms of the 1980s, about the massification of university education, neoliberalism, the international student cash cow, managerial bloat and consultancies —and these are all important factors. But the use of AI in universities has a more immediate history in the choices made by university leaders over the last decade or so. The AI model in universities isn’t an aberration; it’s the logical endpoint of a system built around online content delivery and scalable credential dispensing.
By around 2012, Australian universities had embraced the ‘thru-put’ model: big-box, drive-thru, on-demand, in the cloud, flexible. As a system, it is designed to smoothly pump through as many students as possible, not to provide a better education to a greater number, but because it is both profitable and scalable. Policy backing for the thru-put model dates back to the 2008 Bradley Report, which set the federal government on track for treating twenty-first-century higher education as a black box that produces a mass of skilled graduates by virtue of not much more than enrolment and ‘participation’.
There’s also a strong cultural licence in play that implicates all of us, teachers and students. The thru-put model has thrived because we are okay with this, because those who taste its products all agree to look past its ersatz aftertaste, and keep pretending it’s equivalent to what it blandly simulates.
Inside the field generated by the thru-put model, bot lecturers and bots who do your homework for you pop up as an appealing step-change in efficiency, led and controlled by the universities supplying it. This vision, and the ability to control it, held until about 2022. Then, with the advent of truly industrial-strength LLMs by around 2024, currently available to all students free of charge, without any regulation, the system of drive-thru delivery has peaked, with students now able to extract a credential with even less effort than it took to dispense one.
The rough arithmetic of the thru-put model’s wealth transfer racket
The thru-put model is not about knowledge, learning or pedagogy, but rather is a wealth transfer racket, and it works because it’s so scalable while labour costs are trivially low. Learning may or may not happen, and sometimes good learning still does. But it’s a racket because even if it doesn’t work, students are indebted, while the core beneficiaries are always the most handsomely paid, regardless of performance.
To understand how the racket works, we need to look at ordinary local students enrolling in still normal bachelor degrees: for instance, the first-in-family suburban daughters doing criminology or psychology, the same groups Deakin University sees fit to supply with bot lecturers, and for whom Einstein.ai might seem like a great way to scrape through when your boss lumps you with a last-minute shift, you really need that money or you’re already feeling overwhelmed and unsupported.
The way the racket works for this cohort occupying Commonwealth places is straightforward. It’s majority-funded through contributions from students, who end up with debt, in exchange for a credential. These days, that amounts to over $50,000 for a Bachelor of Criminology, effectively an Arts degree hedging its bets on producing a graduate who will go into an enforcement or public service job. Granted, your 21-year-old will get a degree for their $50,000, but how much of that money will have gone to the delivery of services?
In undergraduate social sciences, unlike medicine or dentistry, the cost of delivery is relatively low. With the demand-driven system, which began in 2012, a much larger number of students enrolled in courses, most of which were increasingly run online. This period was a veritable gold rush for universities. Although today they are worried about post-COVID enrolment numbers and structural decline in such programs, the worry exists not for the students but rather because of the system’s dependence on profitability.
On the best estimates I have been privy to—and it is impossible to get accurate figures because they’re a closely guarded secret—these students get somewhere between 15–30 cents on their dollar. As I calculate it, 70–85 per cent of what is called tuition, which students incur as debt, does not go to the delivery of their education. Monies for tuition, which become student debt, were transferred into and distributed across the university, mostly to staff other than teachers, for activities other than teaching, including consulting, advertising and a vast and growing array of administrative ‘services’, overpriced enterprise software, and ‘open access’ fees to the big academic publishers. As John Ross notes, you’d think this would produce a kind of fattened complacency. But rather, the picture seems to be one of reactivity and panic on the administrative side, as this style of administration, always ramping and bloating, is unable to contain its own costs.
This, in turn, has only strengthened the impetus to lean further into the thru-put model. Minors, niche subjects and electives have all been cut because they don’t produce the super profits now required to fuel the system. And as the vast increase in university staffing since the institution of the thru-put model has been administrative, the lion’s share of wages goes to administration, bureaucrats and the executive. In practice, most of what is paid for by students is a donation to a bloated third entity called ‘the university’, which spends the bulk of that money on its best-paid older incumbents and the prospectus-view aggrandisement that markets the university as a brand, while enforcing austerity, restructuring and wage theft on those least able to fight back: junior academics, sessional staff, librarians and cleaners.
Consider the following rough arithmetic on a big-box thru-put course. Say 300 students enrol in a standard twelve-point social sciences course—all onshore, Commonwealth-supported places. At $2174.88 per student, a course with 300 students brings in $652,464. A course with 460 students brings in a cool one million.
Before the demand-driven system coincided with the thru-put model, healthy numbers in those courses would have been 80–120 people, staffed by an ongoing senior staff member, even a professor. There’d be a few casual tutors, typically PhD candidates, who sometimes got a guest lecture or two. But since 2012, courses with 200 to 400 students are not unusual. In the mid-2010s, at some universities, first-year courses routinely had over 1000 enrolments. Now these big-box thru-put courses are usually run by one coordinator managing two to four sessional staff, who usually only do marking. This is because the coordinator is also a one-person army, who does all the lectures and all the tutorials (sometimes at multiple campuses and online), as well as a big chunk of the marking (80–140 students). Under those conditions, what does academic labour cost?
Assuming a senior lecturer on $150,000 with 500 hours allocated, you’re looking at a teaching expenditure of $37,500. But senior lecturers and 500 hours are plush conditions by contemporary standards. It’s more likely that the coordinator is an associate lecturer on a fixed-term contract earning $80,000 a year, about $40 an hour. Given 300 hours to run the whole subject, the implied labour cost is as little as $12,000. A PhD candidate who sessionally marks 100 students might earn $5–$6000. Multiply by two (assuming the coordinator marks 100 essays), and sessional labour costs in a 300-student thru-put course are probably $12,000. So a course that raked in $652,464 might have cost just $24,000 to run, but assume the range is $24,000 to $50,000, to account for staffing variables. If it got as high as $50,000, assume the school manager would be aggressively looking for ‘efficiencies’ to bring it in line with a dozen other courses costing $24,000.
On those figures, as little as 4.14 per cent of what students paid as their contribution to tuition went on wages for the people employed to deliver the unit. Even allowing for more senior staff allocated more hours, it’s still less than 10 per cent. And if the course was entirely online, as it often is in the 2020s, although there are still high costs (enterprise software, journal access, administration), online students pay for their own buildings, electricity and internet to study from home. So, under the thru-put model, students might routinely pay a tithe of more than 90 per cent.
Even before bot lecturers, then, we can conclude from this rough arithmetic that students don’t pay all that money for an education. Tuition is a misnomer. Students are making a huge donation to that third entity, ‘the university’, whose dying campuses fewer students now visit. They enrol, and they pay, and under these circumstances, the rolling out of bot lecturers is the next logical step.
As of writing, student debt in Australia is more than $80 billion. In November 2024, the federal government took the salutary step of reducing the average student debt by 20 per cent. But it’s questionable whether the remainder of this old debt should be borne by young individuals; they didn’t derive most of the benefit from it. Onshore students studying dentistry or medicine may have had the bulk of their costs paid for by cross-subsidies and taxes. But if students did nearly all big-box thru-put courses, where only 4–10 per cent of that dollar went on delivery, then 90–95 per cent of that $80 billion was transferred to the university to secure the credential. Why should young people, those least able to afford it, pay for this?
The preclusion of excellence and indifference to the absence of competence and bodied experience
The thru-put model, then, is by now an online manifestation of a structurally unfair system, a wealth transfer racket whose business is intergenerational theft, because the youngest, the poorest and the most recently migrated have given nearly all that money to older, wealthier, already advantaged and insulated representatives of the institutions, and to the institutions themselves which now operate as seemingly autonomous systems of growth and accumulation. In keeping with this insight, we can now ask: what does the university produce for students? What value does it create or express for them?
Since 2012, university costs have ballooned and graduate outcomes have declined: the thru-put model is costing more while delivering less to many of the people who pay more for it. Online scalability should have crushed overheads through economies of scale; services should at least be cheaper. But they’re not, because, as sketched above, the thru-put model is highly extractive—and needs to be—to perpetuate, entrench and expand to keep pace with its own swelling.
This flipped relation of the university towards profit rather than education supports one of the craziest values of the thru-put system: its virtual preclusion of excellence, and moral indifference to its absence. Not only does the thru-put model extract as much as 90 per cent of tuition fees from students for the university while enforcing austerity on academic staff, but it also fails to reliably and demonstrably teach and train competently. It can do this because the thru-put model is body-less, and this has allowed its avatars and beneficiaries to peddle a set of false equivalences—often using the language of equity.
Wherever we deal with the body, we demand and expect excellence from competitors and intentionality from participants: we know it’s nonsense to expect results without attention, effort, discipline and care. In elite sport, if we don’t make the cut, we have no place in the program. Where we engage as participants, the bar is set much lower. Anyone can join Park Run. But even if they’re being pushed in a stroller, if they’re involved, they must have shown up, because only bodied effort yields competent participation. If you didn’t show up, you didn’t do Park Run.
Yet the moment we enter the field of higher education under the thru-put model, the demand for excellence is not only dropped—it’s not allowed. Overwhelmingly, this has been justified in the language of access and equity. With disability support and recent awareness around intersectional challenges and neurodivergence, we do have substantive reasons, developed in good faith, to provide learning environments that promote dignity and autonomy, are more accessible and less discriminatory. But a core problem is that the thru-put model is based on high, smooth circulation. Like fossil fuels moving through refineries, it requires continual thru-put fed in at scale, and so doesn’t care overmuch who or what comprises it, as long as it keeps coming. This is one reason why students are routinely referred to by administrators as equivalent full-time student load (EFTSL). I witnessed an Open Day pep talk in which a high-level administrator referred to students as ‘load’, encouraging academic staff to ‘go out and get load!’ On Open Day, parents and students think about hopes and dreams and the future. But the administrators of the thru-put system look at this mass of young people, and they think about load.
For administrators who think of students as load, the language of access and equity has been a huge boon. In lockstep with the managerialist techniques that doubled or tripled academic workloads in the 2010s, access and equity have been used to insist on lowering or removing prerequisites and entry requirements, doing away with staff-to-student ratios, precluding caps on enrolment numbers and removing hurdle requirements for attendance and participation. So administrators duly got load for the university, and load was fed thru, heedless as to whether load learned anything much, attended or consistently tried.
When some academic staff remonstrated about the foreseeably dire effects on pedagogy, classroom dynamics and campus atmosphere, access and equity were invariably the counterargument. The argument for online delivery, which went alongside this, is that outcomes do not depend on presence. In the thru-put model, the value of learning is accessible content, so intellectual engagement can occur anywhere. At the Australian Institute of Sport or any local football match, body-less engagement would simply be rejected as nonsense. At the thru-put university, accessible content is the greatest good and digital accessibility is the key, without the need for in-person learning or teaching. The argument that clinches this deal is that insisting on the attendance of actual bodies discriminates against those who do not attend. Anti-discrimination is offered back as a value that cannot be contested.
The removal of hurdle requirements is seldom seen as the huge problem it is. The basic expectation that anyone would or should actually attend university in person, or attend a certain number of lectures or tutorials, is removed.
Abolishing hurdle requirements hides a number of other false equivalences that are also represented as values.
Once you get rid of hurdle requirements, you allow that the student who doesn’t engage or show up is just as much a student, and just as passable, as somebody who did try, who did show up, week after week. The moment you do this in a course of 300-plus eighteen-year-old strangers, you birth a game logic where the authentic and diligent student is always the chump, or just someone eccentric who persists in doing something idiosyncratic even though it’s objectively not in their best interest. And you are doing this to a generation of people raised on gaming, for whom gamification is breathing. The thru-put university has modelled its own truth: consistent attendance and heartfelt effort don’t matter: ‘work smarter, not harder, embrace AI’.
Removing hurdle requirements puts a penalty on good faith and effort: it’s a tax on heart, a devaluation of spirit. It actively incentivises not showing up, because when you do bother to roll in, you realise that attending isn’t the vibe, because most of your classmates aren’t there. And indeed, if everything is recorded and on Canvas (or Moodle), only two to three pieces of digitally generated, digitally submissible assessment are needed to pass, and you can get 51 per cent in each and still be deemed passable and still get the same degree, what more is there to university?
Unsurprisingly, under the thru-put system, pedagogical culture shifted to being overwhelmingly about assessment and anxiety about its instructions, rubrics and marks. For teaching staff, much of the work is no longer about a course’s substantive topic. Post-2012, a huge amount of ‘teaching’ labour goes into dealing, sincerely, with small uncertainties that mean a great deal to students, because only the final number (mark) has value, and students are paying so much. A lot of this involves micro-arbitrage over marks and comments: ‘why did I get 76, not 78?’; ‘can you bump my 78 to an 80?’ But now, there’s a bot to spare teaching staff this stressful email reply. Already, in a number of programs, students are being graded by AIs; no remarks, no recourse, full price.
For many students under the thru-put system, going to university has largely come to mean logging in, navigating the behaviourist nudges and marking rubrics of gamified learning environments, and dealing with incipient resentment and the chronic anxiety that comes with not getting timely or meaningfully human engagement from the ‘learning management system’. It has to be said that perhaps 20–30 per cent of students do well in these smoothed online platforms, but another large group feel confused, alienated and betrayed. Often, the online teaching materials were written years ago by overworked sessionals. Or they never made good sense, because they weren’t thoughtfully designed and don’t measure competence, skill or excellence that accurately, and no one has revised them because everyone is too busy, and a fungible system has no cultural memory.
It’s important to note that it’s the thru-put model that has also led to the near death of universities as physical places of bodied, lively, on-campus experience. The absence of hurdle requirements and face-to-face learning systematised by this model means that most ‘university students’ do not undergo any kind of experiential process ‘at a university’, nor do they come to understand its value or what it might have been. It’s just not there anymore. Online, there are no buildings, but there is also no bildung, no rites of passage, no rituals, nothing to be involved in or part of; nothing to remember, no youthful follies to look back on. An older generation remembers campus; can anyone think fondly of Canvas?
As with equity and diversity, the fate of ‘experience’ at our universities has also been a devaluation of meaning in defence of the thru-put model. Students who go to university may no longer go to their university, but some university executives talk obsessively about ‘the student experience’. What they mean by this is something more like better UI/UX: the idea that students should have a smooth simulation, with smooth onboarding, a dashboard that gives them seamless access to content on their mobile device, and a smooth progression through their program, all the way to graduation. When they graduate, they can don regalia and come and have a photo taken in front of some Hogwarts-y building, thereby completing the simulacrum of the experience.
This is where we are in 2026. University education in Australia has largely become a garbage online experience—smooth garbage perhaps, but with rough, garbage real-world outcomes, both pedagogically and socially. For hundreds of thousands of young people, the real student experience is that of an anxious and lonely nineteen year old in a gaming chair in their suburban bedroom who doesn’t know their teacher/instructor or any of their cohort, and who is rinsing their work through GPT, Claude or Gemini the day the assessment is due, because in their heart, they have no confidence in their skills and knowledge. Thanks to the thru-put model, Australian universities are now digital McDonald’s without Happy Meals.
A way beyond the thru-put model?
The thru-put model was not imposed from outside: it was facilitated by federal policy, built by university leaders, and has been sustained by our collective willingness to accept the simulacrum. An institution deploys a bot lecturer, and a student deploys a counter-bot essay writer. None of the intended learning has occurred; teaching staff are burned out; no one is accountable; the credential is still issued; the debt still accrues; the university executive still gets paid, gets load and spruiks the online ‘student experience’. ‘Quality without compromise.’ ‘It’s never been a better time to take your career to the next level.’
But what is higher education actually for, and what would it take to choose to build something fit for purpose, given the pressing, mounting local and global issues we must face up to soon? One thing is certain: neither the thru-put model’s avatars and beneficiaries, nor its simulated experience, will offer a fertile answer.
Comments
This is a devastating – so sadly too, a true description of so much of what is occurring in the sector, for which the author deserves high praise. Academics who fight for hurdle requirements are greeted with arguments from equity that empty classrooms and degrade the educational experience in ways almost unfathomable three decades ago. AI is making the incentive structures worse, as educators now have to compete with supercomputers, while universities seem – amazingly, given their history – to be largely accepting of AI and even trying to welcome it into classrooms. The situation at times evokes absurdist literature, so bizarre are its features.
Thanks for the devastating critique of universities today, which demonstrates how capital has become detached from labour. What’s the alternative? Smaller colleges, such as those found in the USA, might foster greater integrity. Perhaps even knowledge cooperatives, where staff have a stake in the institution. Raewyn Connell’s The Good University: What Universities Actually Do and Why It’s Time for Radical Change is worth a look.