Australia has spent the past two years scrambling to write rules for artificial intelligence in classrooms, lecture theatres and exam halls. Far less attention has gone to the part of the education system where children are youngest and least able to speak for themselves. A group of Australian experts now says that gap needs closing, and that the answer is not to simply copy and paste the policies written for schools and universities onto childcare centres and kindergartens.
That is the central message of new analysis reported by Bioengineer.org, in which Australian specialists argue for a cautious, sector-specific approach to AI in early childhood education and care. Rather than one broad national rulebook covering every learner from toddlers to postgraduates, they contend that the under-fives deserve their own carefully bounded framework, one that reflects how very young children actually learn and how little say they have over the technology placed in front of them.
Why early childhood is different
The reasoning turns on a point that sounds obvious once stated but has largely been missing from the national conversation. The tools, risks and trade-offs that apply to a Year 10 student drafting an essay with a chatbot have almost nothing in common with those facing a four-year-old in a long day care room. Older students can be taught to question an AI system, to check its output and to disclose when they have used it. A preschooler cannot. That places the entire burden of judgement on educators, providers and the parents who hand over consent on a child’s behalf.
There is also the question of what AI is even for at this age. In early childhood settings, the technology is far more likely to sit in the background, helping with administration, documentation, observation notes, rostering and communication with families, than to be put directly in a child’s hands. The experts’ argument is that policy should be written for those real uses, with clear limits on data collection, surveillance and any system that profiles or assesses children, rather than borrowing assumptions from a schooling context where students interact with the tools themselves.
Privacy sits at the heart of the caution. Early learning services already gather a remarkable amount of information about children and their families, from developmental milestones and photographs to health records and attendance patterns. Layering AI-driven analytics or cloud-based platforms on top of that raises the stakes considerably, because the people being observed are too young to understand, let alone object. A sector-specific policy, the argument runs, would set guardrails on exactly this kind of data use before it becomes routine.
Two ways to read the warning
The cautious camp sees sequencing as the whole point. Get the safeguards right first, they say, and the benefits can follow without exposing children to harms that only become visible years later. On this view, a rushed, generic framework is worse than no framework at all, because it lends official cover to practices that were never designed with toddlers in mind.
There is a competing instinct, though, and it is worth taking seriously. The early childhood sector in Australia is under genuine strain, with chronic workforce shortages, high turnover and educators spending large slices of their day on paperwork rather than with children. Used carefully, AI that trims administrative load could hand time back to the people doing the work, which is precisely the outcome everyone says they want. From that angle, an overly restrictive or slow-moving policy risks locking the sector out of tools that could ease the pressure, at a moment when it can least afford to lose staff.
The two positions are not as far apart as they look. Both accept that some uses of AI in early learning are reasonable and some are not. The disagreement is about pace and about who bears the risk if the settings are wrong. A sector-specific framework is, in effect, an attempt to let the sector capture the upside on administration while drawing a firm line around anything that touches children directly.
What it means for Australia
Australia already has form here, just not for the youngest learners. The Australian Framework for Generative Artificial Intelligence in Schools, agreed by education ministers, gave the schooling system a shared set of principles covering safety, privacy, transparency and fairness. Higher education has had its own running debate about assessment and academic integrity. Early childhood, governed through the National Quality Framework and overseen by the Australian Children’s Education and Care Quality Authority, has largely been left to work it out service by service.
That matters because early learning is also squarely in the federal government‘s sights. Canberra has committed to major reform and investment in the sector, including moves toward cheaper and more universal access, which means more children, more services and more data flowing through more systems. Decisions about AI governance made now will shape how that expanded system handles information about hundreds of thousands of Australian children. Getting ahead of the technology, rather than reacting to it after a problem surfaces, is the practical case for acting before adoption outpaces the rules.
For providers, the immediate takeaway is that the absence of a dedicated framework is not permission to do whatever the latest platform allows. For families, it is a reminder that consent given at enrolment may be carrying more weight than anyone intended. And for policymakers, the experts’ intervention is a nudge to treat early childhood as a distinct domain with its own standards, not an afterthought bolted onto schools policy.
What’s next
The call is unlikely to produce an overnight response. Any sector-specific framework would need to be reconciled with the National Quality Framework, with Australia’s broader work on AI guardrails and with the privacy reforms moving through at a national level. The more probable path is a period of guidance and consultation, with peak bodies, regulators and researchers shaping what responsible use looks like before anything binding lands. The experts’ point is simply that the clock is running: AI is already arriving in early learning services whether or not the policy is ready for it, and the youngest Australians are the least equipped to manage what comes with it.
Sources: Bioengineer.org, via GNews.
