Every few years a new technology arrives in Australian classrooms promising to lighten the load, and every few years the same awkward question comes back around. If the tool is doing the heavy lifting, what exactly is the student doing? That is the thread Professor Martin Westwell, chief executive of the South Australian Department for Education, pulls on in his latest column for Teacher Magazine, and it lands at a moment when schools across the country are trying to work out where generative AI belongs in a child’s education.
Westwell’s argument is deceptively simple. Explicit instruction, classroom technology and artificial intelligence are usually treated as separate debates, each with its own camp of believers and sceptics. He puts them under a single lens. In each case, he says, the question worth asking is not whether a method is modern or traditional, hands-on or teacher-led, but who is carrying the cognitive effort at the point where learning actually happens. The point of good teaching, in his telling, is to make sure the demanding mental work is landing on the student rather than being quietly outsourced to a worksheet, a screen or a chatbot.
The thinking you must not take away
Learning, as cognitive scientists have long argued, is a by-product of thinking. A student remembers what they have had to grapple with, not what has been handed to them in finished form. That is the principle sitting underneath Westwell’s warning. When a tool removes the struggle entirely, it can also remove the learning, even as it makes the lesson look smoother and the output look more polished.
“The aim is not to make learning harder,” Westwell writes in the column. “It is to make sure the hard thinking is happening in the right place.” It is a line that cuts against a lot of the marketing around education technology, which tends to sell friction as the enemy. His position is that some friction is the entire point, and the craft of teaching lies in knowing which friction to keep and which to clear away.
This is where his treatment of explicit instruction becomes interesting. Explicit instruction, with its emphasis on clear modelling, worked examples and carefully sequenced practice, is often caricatured by critics as spoon-feeding. Westwell flips that reading. Done well, he argues, it manages a student’s cognitive load so that mental effort is spent on the ideas that matter rather than wasted on confusion, which frees the student to do the genuinely hard thinking rather than sparing them from it. The same logic, he suggests, should govern how teachers reach for technology.
Two ways to read the AI question
On one side are the enthusiasts who see generative AI as a tutor in every pocket, capable of explaining a tricky concept five different ways and never losing patience. There is real substance to that case. A tool that can scaffold a struggling student, draft a first attempt they then critique, or generate endless practice problems is a tool that can extend what one teacher can do for thirty kids. Used deliberately, it can put more thinking in front of more students.
On the other side sit educators who worry that the convenience is the trap. If a student can produce a passable essay, a solved equation or a tidy summary without having understood any of it, the assessment looks fine while the learning never happened. That concern is already reshaping how Australian institutions think about marking and oversight, a theme FluentSea has tracked in reporting on the University of Tasmania‘s move to rethink assessment around AI, and in the arguments of educators such as HP’s Brett Salakas, who has made the case that AI in schools should push lessons to be more human rather than less. Westwell’s framing does not resolve that split so much as give both sides a shared test. The question is not whether AI is in the room. It is whether the student is still the one thinking.
What it means for Australian schools
The timing matters here. Australian education ministers have signed off on a national framework for generative AI in schools, and systems from South Australia outward have been running trials of AI tools built specifically for classroom use. South Australia itself has been among the more forward-leaning jurisdictions, piloting a schools chatbot designed with guardrails rather than banning the technology outright. That makes Westwell’s intervention more than an academic reflection. He is not a commentator on the sidelines, he runs one of the country’s larger school systems, and his department’s choices ripple through the lives of hundreds of thousands of students.
For teachers, the practical takeaway is a design principle rather than a policy. Before reaching for any tool, AI included, the question becomes where the thinking will end up. A chatbot that drafts feedback a teacher then refines keeps the professional judgement with the human. A chatbot that writes the student’s answer removes the very effort the task was meant to create. The tool is neither good nor bad in itself. What matters is the shape of the thinking it leaves behind.
There is a workforce dimension too. Australia is already wrestling with teacher shortages and workload pressure, and AI is frequently pitched as relief, taking lesson planning, admin and marking off educators’ plates. Westwell’s logic applies just as neatly there. Automating a teacher’s busywork so they have more time for students is a clear win. Automating the professional thinking that makes a teacher a teacher is a different proposition altogether, and the line between the two is not always obvious in a vendor demo.
What’s next
The harder work is turning a principle into practice. Deciding which thinking to protect is a judgement call that will look different in a Year 2 reading lesson and a Year 12 physics class, and it is the kind of call that resists a one-size-fits-all rule handed down from a department. Expect the national framework to keep evolving as evidence comes in from the trials now running, and expect assessment to remain the sharp end of the debate, because that is where the gap between looking competent and being competent shows up first.
Westwell’s contribution is to give crowded staffrooms a single, portable question they can carry into any decision about any tool. Strip away the hype and the backlash, and the test is the same one good teachers have always applied. Who, in this lesson, is doing the thinking? Keep the answer pointed at the student, and the technology can be a genuine ally. Lose sight of it, and even the cleverest tool becomes a way to look like learning is happening while it quietly stops.
Sources: Teacher Magazine, ACER
