Generative AI arrived in classrooms and lecture halls faster than any curriculum, assessment policy, or teacher training program could reasonably keep pace with. Schools and universities are now running a live experiment on what education is for, in real time, with today's students as the test cohort.
Why this matters right now
Every Australian school and university has had to make fast, imperfect decisions about AI: ban it, embrace it, or something in between. Assessment integrity is the sharpest edge of the debate — if a student can generate a competent essay in seconds, what exactly is being assessed, and how? But the deeper question is about capability, not compliance: what should young people actually know and be able to do when a huge share of routine intellectual work — drafting, summarising, coding boilerplate, first-pass analysis — can be done by a machine in seconds? The answer changes what's worth teaching, not just how to catch students who cut corners.
There's also an equity dimension. Students, schools and families with better access to good AI tools, and the guidance to use them well, stand to gain the most from them; students without that access, or without the foundational literacy and numeracy to use AI critically rather than blindly, risk falling further behind. AI could narrow gaps or widen them, depending entirely on how deliberately the education system responds.
The central tensions
- Foundational skills vs tool fluency. Does heavy reliance on AI in learning erode the underlying skills — writing, calculation, reasoning — that AI itself depends on being taught in the first place?
- Assessment integrity vs assessment relevance. Is the priority stopping AI-assisted cheating, or redesigning assessment so the question of "cheating" barely applies?
- Equity of access vs equity of guidance. Giving every student the same AI tool doesn't guarantee the same outcome if only some students are taught to use it critically.
- Teacher workload vs teacher relevance. AI can reduce administrative burden on teachers, but only if it doesn't simultaneously reduce the perceived value of teaching itself.
Questions to bring to the discourse
- What should a school-leaver in 2030 be able to do that a good AI model cannot do for them?
- Should generative AI use be taught explicitly as a skill, the way typing or research skills once were?
- How do we stop AI in education from becoming another axis of inequality between well-resourced and under-resourced schools?
- Is the current wave of AI detection and integrity policy solving the real problem, or just managing its symptoms?
In one sentence
AI hasn't just changed what students can do — it has reopened the much older question of what school and university are actually meant to achieve.