There is a reflex, whenever a new technology lands in schools, to assume it will squeeze the humanity out of the place. Calculators would rot mental arithmetic, the internet would end reading, and now artificial intelligence, we are told, will hollow out the relationship between a teacher and a child. A new episode of the Education Matters Australia podcast sets out to test that assumption head on, asking whether AI might do the opposite and hand classrooms back some of the human connection that administrative overload has quietly eroded.
The guest making that case is Brett Salakas, a former primary school teacher who now works on education strategy at HP and has spent more than a decade as one of the more recognisable voices in Australian edtech. He is the founder of the long-running #aussieED community, a network of teachers who swap classroom ideas online, and he has built a reputation for being enthusiastic about technology without losing sight of what actually happens in a room full of nine-year-olds. That background matters, because it shapes the argument he brings to the episode: technology in schools is only worth anything if it serves the people in front of it.
The pitch: give teachers their time back
The central idea Salakas puts forward on the podcast is that the most valuable thing AI can offer a school is not a smarter student or a flashier lesson, but time. Australian teachers routinely report working well beyond their paid hours, with lesson planning, differentiation for students of varying ability, report writing and endless administrative paperwork eating into evenings and weekends. If generative tools can take a first pass at a rubric, draft a parent newsletter, or sketch three versions of a worksheet pitched at different reading levels, the argument goes, the teacher is freed to spend that reclaimed time where machines cannot help, which is sitting beside a struggling child or noticing the kid who has gone quiet.
That framing, of AI as the thing that does the drudgery so humans can do the human part, is the thread the episode pulls at. It is a deliberately optimistic reading, and the conversation, which you can hear in full via Education Matters Magazine, leans into the possibility rather than the panic.
A more cautious view from the sector
Not everyone in Australian education shares that confidence, and the counter-argument deserves equal weight. Teacher unions and many academics have warned that the time AI supposedly saves can just as easily be clawed back by employers who load up expectations, and that the quality of an AI-drafted report is only as good as the teacher’s capacity to check it. There is also the awkward reality that students are using the same tools, often more fluently than their teachers, which has already forced universities and schools to rethink how they assess learning at all. The University of Tasmania, for instance, has been publicly wrestling with how to keep human oversight in assessment as AI use spreads, a reminder that the technology cuts both ways inside a single institution.
The federal framework governing this is still catching up. The Australian Framework for Generative Artificial Intelligence in Schools, agreed by education ministers, sets out principles around safety, fairness, privacy and transparency, but it is guidance rather than hard regulation, and implementation is left largely to states, systems and individual schools. That gap between a high-minded national framework and the messy reality of a term-four classroom is exactly where the debate Salakas joins actually lives. Enthusiasm from a vendor-aligned voice will always be read, fairly or not, through the lens of commercial interest, and HP sells the devices these tools run on. The episode does not pretend otherwise, which is part of what makes the conversation worth having in the open.
Why this matters for Australia
For Australian schools the stakes are not abstract. The country is in the middle of a genuine teacher shortage, with workforce pressures acute in regional and remote areas and in subjects like mathematics and science where qualified staff are hardest to find. Anything that credibly reduces the hours a teacher spends on paperwork is not a nice-to-have, it is a potential retention tool in a profession bleeding experienced people. At the same time, Australia’s deeply unequal school funding landscape means the schools most able to buy devices, train staff and roll out AI thoughtfully are often the ones that already have the most, raising the risk that a technology pitched as a leveller quietly widens the gap.
There is also a distinctly Australian question of digital equity across the city and country divide. A tool that saves a Sydney teacher two hours a week is worth far less to a colleague on a patchy connection in the bush, and national rollouts have a long history of stumbling on exactly that terrain. If the promise of more human classrooms is going to be real for every child and not just the ones in well-resourced metropolitan systems, the infrastructure and professional development have to arrive alongside the software, not years behind it.
What comes next
The honest answer coming out of conversations like this one is that nobody yet knows whether AI will make schools more human or less, because it depends almost entirely on decisions that have not been made. Whether the time saved goes back to students or gets absorbed by new demands, whether teachers are trained to supervise these tools or simply handed them, and whether access is equitable or lopsided, all of that will be settled school by school and system by system over the next few years.
What the podcast gets right is the reframing. The useful question is not whether AI belongs in the classroom, because it is already there in students’ pockets, but what we want teachers freed up to do with the time it gives back. If the answer is genuine human connection, then the technology is doing its job. If the answer turns out to be more paperwork in a different format, the sector will have learned nothing from the last wave of promises. For a country short on teachers and long on good intentions, getting that answer right is the part no algorithm can do for us.
Sources: Education Matters Magazine.
