Opinion
What is the point of opening the doors if nobody is paying attention to what happens once students are inside?
The Universities Accord (Opening the Doors of Opportunity) Bill cleared parliament on 13 October, the seventh piece of Accord legislation. It introduces managed growth and needs-based funding, defers to the Australian Tertiary Education Commission (ATEC) the job of allocating places, and promises 230,000 extra student places over the next decade.
The Minister's release announcing the bill doesn't mention teaching, learning, curriculum, or assessment. Nor does the Universities Australia release welcoming it, which mentions teachers only as a workforce shortage.
That is a fair summary of the national conversation for the last five years, going back through the Accord process. The debates have been about places, caps, and funding formulas; international student numbers; governance, mergers, and executive pay; and campus safety, racism, and antisemitism.
Much of this matters, and some of it is long overdue. Students should be safe and should feel they belong. But a student can be safe, supported, and satisfied, yet still leave having learned very little, if anything, of relevance.
It is an open secret that very few students are actually completing the volume of learning required. Many are engaged in learning well below the number of hours assumed in student load (in our data – from anonymous institutions – the mean is half the hours). What, then, have we opened the doors to: a system where the core process is never discussed and only half as much of it is happening as is supposed to be?
ATEC's consultation paper on its priorities for 2027–28, released this month, doesn't change the picture. It adds 'student experience and safety' as a new priority. In a paper setting nine priorities, 'assessment' doesn't appear once and 'curriculum' appears twice, both in passing. AI earns two mentions, one of which promises to 'deliver teaching more efficiently' (whether that is even possible is entirely debatable). Academic staff need only be 'well prepared to facilitate good teaching' (without any clear sense of what that means).
Student experience and safety are important. So is ensuring students have adequate support and flexibility to meet the challenges they increasingly face outside their studies. However, at some point the sector needs to focus on whether it is effectively delivering on what it says on the tin (i.e., a ‘higher’ education). An overdue discussion about whether people teaching in the sector need to be qualified to do so (seems a no-brainer to me) is not enough.
The data suggest we should be paying much closer attention. The 2023 cohort, the first to start university with generative AI and AI agents, recorded the lowest first-year attrition in a decade: 12.2% of domestic bachelor students didn't return, down from 14.7%. Success rates reached a ten-year high of 87.9% in 2024.
The Department's explanations, including Fee-Free TAFE, a strong labour market, and a younger cohort, are reasonable, and nobody can say for certain that AI caused the improvement. The problem is that nobody can say it didn't. The system has no reliable way of telling students who are actually learning more from students who are passing more. This is the performance paradox of AI writ large.
We have been fooled by a good first-year retention number before. Andrew Norton showed that fewer of the 2020 cohort left after first year, but four-year attrition then reached 23.8%, the highest since tracking began in 2005.
AI offers a new route to the same deferral. In one widely cited trial, high school students with unrestricted access to GPT-4 did 48% better on practice problems and 17% worse than a control group once it was taken away.
Despite a colossal effort across the sector, much assessment is still unsupervised writing, and fluent output makes it easy for students to overestimate how much they have learned. Some will find out or get found out in second or third year, with HELP debt and no qualification.
ATEC's own draft outcomes have students 'finishing in minimum time incurring minimum debt', though the aim of having students invest minimum time seems already achieved. Students using AI to do the work increasingly risk neither finishing nor avoiding accumulating unnecessary debt.
The students most exposed are the ones the Accord most wants to recruit; those admitted through non-ATAR pathways are already nearly twice as likely to leave in first year. These are the same cohorts who are most vulnerable to learning loss from the introduction of AI.
The point of all this reform must be graduates who have met a program's learning outcomes and can flourish in a world where AI is ubiquitous. ATEC's consultation asks whether there are key focus areas within its remit that 'should be added'. Taking seriously what happens, and what ultimately leads to students completing, once students walk through the door (i.e., adaptive learning for the age of AI) would be a good place to start.
There is no point in creating an amazing higher education experience if students don’t learn anything. If we are not willing to have the difficult conversations about learning in the age of AI, we may as well give up the charade, turn our campuses into entertainment precincts, and be done with it.
Professor Jason Lodge is Director of the Learning, Instruction, and Technology Lab at The University of Queensland