The AI future of tertiary education is being decided in five arguments: whether assessment can still certify learning, what the graduate workforce will need, whether personalised learning is real or a sales pitch, whether credentials keep their value, and whether regulation can keep up. Of those five, only two will change practice in Australian providers within five years, and they are the ones TEQSA has already written about.
This article sets out each debate as it is actually being had, and gives my view of which ones will move from conference panels into board papers. It draws on fifteen years of TEQSA registration and governance work, most of it with private providers who cannot afford to bet on the wrong argument.
Which debates define the AI future of tertiary education?
The debates are not new, but the tools are. Generative AI arrived in late 2022 and by mid-2023 the regulator had published Assessment reform for the age of artificial intelligence, a short paper built on two principles: that students must be equipped for a society pervaded by AI, and that trustworthy judgments about learning require multiple, inclusive and contextualised approaches to assessment. That paper reframed the discussion. It moved the question from "how do we stop students using AI" to "how do we know a graduate can do what the testamur says".
The other debates followed from that reframing. If assessment changes, what a degree certifies changes with it, which is the credential debate. If graduates enter workplaces where AI drafts, codes and analyses, what should the curriculum teach, which is the workforce debate. And if the regulator is asking providers to redesign assessment at program level, is it acting inside the Threshold Standards or ahead of them, which is the regulation debate.
The assessment debate is already settled in principle
In my experience this is the one debate that has moved past argument. The sector has converged on a pattern: assurance at program level, with a small number of secured assessment points where the provider can be confident the work is the student's own, surrounded by tasks that assume AI is available and assess how well it is used. The label "two-lane" is a sector coinage rather than TEQSA's, but the substance is what the regulator described in its 2025 follow-up, Enacting assessment reform in a time of artificial intelligence.
What remains unsettled is execution. Redesigning assessment across every accredited course is slow, and it collides with staffing, learning resources and the academic board's approval cycle. Providers that treated the June 2024 request for a credible institutional action plan as a one-off submission are now finding that TEQSA reads the plan against what has actually been done since. I discuss the practical side in conversations on AI in higher education.
The workforce and personalisation debates will change slowly
The workforce debate is loud and, for most providers, premature. Employers say they want graduates who can work with AI; they cannot yet say what that means for a nursing, accounting or design curriculum. Course learning outcomes under Standard 1.4 must be specified and assessed, and a provider that writes "AI literacy" into every outcome without being able to assess it has made its application weaker, not stronger. I expect the professional accreditation bodies to settle this discipline by discipline over the next five years, and providers should follow rather than lead.
Personalisation is the debate I trust least. Adaptive platforms have promised individualised learning for two decades, and the evidence that they improve attainment at scale remains thin. AI tutors will improve student support, particularly for providers with small cohorts and limited staff, but they do not change what Standard 3.3 requires: learning resources and educational support that are actually delivered and monitored. I set out the sector positions more fully in the role of AI in the tertiary education dialogue.
The credential and regulation debates are where boards should look
The credential debate is the sleeper. If an assessor cannot tell whether a graduate met the learning outcomes, the qualification loses value, and that is a Standard 1.5 problem as much as a Standard 1.4 problem. TEQSA's request for information on gen-AI risk was framed explicitly around award integrity, and I expect renewal assessments to keep asking how the governing body knows its awards remain credible.
The regulation debate is often misstated. TEQSA has not written AI into the Threshold Standards, and its papers are guidance, not law. But the Standards are in the present tense and technology-neutral: assessment must be valid, integrity must be protected, and the academic board must be overseeing both. The regulator does not need new standards to ask hard questions about AI, and it is asking them. The mistake is to wait for a rule that is not coming.
Where I think practice will actually move
Two things will change practice in five years. Assessment design will be rebuilt at program level with secured points, because the regulator, the professions and the students' own employers are all pushing in the same direction. And governance oversight of AI will become a standing item, with the academic board receiving reports on integrity incidents, assessment redesign progress and staff capability in the same way it receives progression data now. Our earlier piece on AI considerations in higher education describes what that reporting looks like.
The rest of the AI future of tertiary education, the personalised campus and the reinvented curriculum, will arrive more slowly and less dramatically than the panels suggest. The providers that do well will be the ones that fix assessment and governance first and treat everything else as an experiment with a budget and an end date.
Download the AI Governance Checklist
— a one-page set of questions a governing body and academic board should be asking about generative AI, drawn from our TEQSA registration and governance work with private providers. Get the checklist
Want the full article?
Enter your email for free access to the rest of this guide and our TEQSA resource library.
Frequently asked questions
Has TEQSA changed the Threshold Standards because of AI?
No. The Higher Education Standards Framework (Threshold Standards) 2021 does not mention generative AI. TEQSA's assessment reform papers and its 2024 request for action plans are guidance and regulatory correspondence, not amendments to the Standards, but the existing standards on assessment, integrity and academic governance already give TEQSA the basis to ask about AI.
Does TEQSA require AI detection software?
No. TEQSA has not endorsed detection software as sufficient evidence of assessment integrity. The sector pattern it has described relies on program-level assessment design with secured assessment points rather than on detection alone.
Which AI debate should a private provider act on first?
Assessment. Redesigning assessment at program level, with secured points and tasks that assume AI is available, addresses the risk to award integrity that TEQSA has identified, and it is the area where the regulator has published the most guidance.
Will AI change what a degree certifies?
Probably at the margins, as professional bodies revise what competence means in each discipline. The AQF level descriptors and course learning outcomes remain the reference point, and a provider must still be able to show that graduates meet them.
Dr Brendan Moloney is CEO of Darlo Higher Education, Australia's largest specialist TEQSA consultancy. He holds a PhD from the University of Melbourne, is a Cambridge University Press author on governance in higher education, and has advised private providers on registration and course accreditation for more than fifteen years.