Your Smartest Student Is Not Human Anymore

Somewhere in a Malaysian classroom this morning, a teacher stated a fact, and a student quietly checked it against the machine in their pocket before the sentence was finished.

Sit with that scene for a moment, because everything about the future of teaching lives inside it. For most of the history of schooling, the classroom ran on a simple information economy: the teacher had the knowledge, and the students came to receive it. We even have an affectionate name for the model, the sage on the stage, the central figure who holds what others lack and transmits it in carefully measured doses. It worked, more or less, for as long as its one load-bearing assumption held: that the teacher was the smartest source of knowledge in the room.

That assumption has now quietly died. The smartest student in the classroom is no longer human, and it is enrolled in every class, sitting in every pocket, fluent in every subject on the timetable.

To be fair to our profession, we saw the sage's retirement coming long before AI arrived. As early as 1993, Alison King published her now-famous call for teachers to move from sage on the stage to guide on the side, arguing that the transmission model produced passive learners ill-suited for the century ahead, and that the constructivist alternative, students actively building knowledge while teachers facilitate, was the better bet (King 1993). Decades of learning science backed her. The landmark meta-analysis of 225 studies found that students in traditional lectures were one and a half times more likely to fail than students in active learning classrooms, a result so lopsided its authors remarked that if it were a medical trial, the lecture condition might have been stopped early for causing harm (Freeman et al. 2014). The direction of travel was never in doubt.

Then came the branding. When the world began talking about Education 4.0, the education sector's answer to the Fourth Industrial Revolution, the argument acquired an economic engine: nations wanted agile, adaptive workforces, and that meant classrooms where students co-create knowledge rather than memorise it. Malaysia joined this conversation enthusiastically, and our policy documents absorbed its vocabulary. Notably, artificial intelligence was always on Education 4.0's list of disruptive technologies. But here is the detail we forget: when that framework was drafted, AI was a laboratory promise and an industry tool. It was not democratised. No student had it. The disruption was hypothetical, scheduled politely for the future.

The future arrived off-schedule. The public release of ChatGPT, followed by a plethora of ever-more-capable systems, put what I can only call an augmented intelligence into the palm of every student's hand, free or nearly so, in English and increasingly capable Malay. The hypothetical became a reality. And tellingly, the frameworks began multiplying to catch up: scholars now write of Education 5.0, reframed around human-centricity and wellbeing in the image of Society 5.0, and some have already begun sketching Education 6.0, with systematic reviews noting how young and unsettled this literature remains (Shahidi Hamedani et al. 2024). I confess some wariness about the version numbers; education is not software, and numbering our paradigms can substitute labelling for thinking. But the churn itself is the signal. The field is renumbering itself because the ground genuinely moved.

Consider what a student can now do, unsupervised, before breakfast. Verify any fact a teacher utters, instantly; checking claims that once required a library visit is simply a reflex now. Ask for a complex equation or concept to be explained again, slower, with a different analogy, in simpler language, in Malay, without embarrassment and without exhausting anyone's patience. Rehearse an argument by sparring with the machine, testing claims and counterclaims before ever raising a hand. Generate practice questions, worked examples, and summaries tuned precisely to what they do not yet understand. These are not fantasies from a vendor's brochure. They are the daily, mundane reality of studying in 2026, and any teacher who has not noticed has simply not asked.

So what remains for the teacher when the knowledge monopoly is gone?

Everything that actually mattered, is my answer. The monopoly was always the most replaceable part of the job. What the machine in the pocket cannot do is what I have spent much of this column series describing: it does not know the students, it does not carry pedagogical intent, and its fluent output demands exactly the critical judgement that novices have not yet built. On that last point the research offers a warning we should take personally: heavy, unguided reliance on AI tools is associated with weaker critical thinking through cognitive offloading, particularly among younger users (Gerlich 2025). An augmented student is not automatically a better-educated one. Augmentation without orchestration can produce fluent dependence, and the classroom is precisely the place where that trajectory gets corrected, or does not.

This is why the teachers who will thrive are not those who out-lecture the machine, an unwinnable contest, but those who redesign the room around what the machine cannot do. A dynamic, noisy, lively classroom, students debating, questioning, testing AI outputs against each other and against the syllabus, defending arguments they built the night before with a chatbot's help, should not be feared as a loss of control. Designed properly, and suited to the heterogeneity of the actual students in the room, it is a more productive and more impactful classroom than the silent transmission hall ever was; that is precisely what the active learning evidence has said all along. The constructivist paradigm King championed three decades ago turns out to have been waiting for this moment: a classroom configured for interaction, where students bring machine-augmented preparation and the teacher's craft lies in orchestrating what happens when those preparations collide. The reviews of large language models in education arrive at the same conclusion from the technical side: the benefits materialise when teachers and students bring competence and critical awareness to the tool, and the risks compound when the humans in the room abdicate (Kasneci et al. 2023).

None of this diminishes the teacher. It relocates them, from the stage to somewhere harder and more important: the design desk, the discussion floor, the moments of judgement no machine can make. The sage's stage has collapsed, not because sages failed, but because the economy of scarce knowledge that built the stage is gone. What students need now is not another source of answers. They are drowning in answers. They need someone who knows them, who can configure the room, the task, and the conversation so that answers turn into understanding.

The smartest student in the room is not human anymore. Which is exactly why the wisest person in the room still must be.