Most schools I talk to are still debating whether to allow AI. Their students settled that question months ago, on their own phones, with no supervision and no framework for judging what came back.

That gap, between the policy conversation happening in the staff room and the practice already happening in the classroom, is the real problem, and it is not a technology problem at all.

A secondary school English teacher put it to me plainly: "You know, these days every written assignment I hand out will have some students sending in AI work. There's no way that particular student suddenly became a language expert. Yet they deny using AI, they always do. I try using AI detection tools, but my other students tell me they have found ways to get around it."

That is not a story about dishonest students, or not only that. It is a story about an arms race the school never chose to enter, detection tool against workaround, denial against suspicion, with nobody actually teaching the student what responsible AI use in that assignment would have looked like in the first place.

Access was never the hard part

Give a student a smartphone and an internet connection and they will find a way to use AI for homework, for essays, for exam revision, whether or not a single teacher in the building has decided this is allowed. That much was predictable, and arguing about it now is roughly as useful as arguing about whether students should have calculators. The access question is already closed, whatever a school's official policy says.

What is not closed, and what almost nobody is teaching deliberately, is judgment: knowing when an AI answer is good enough, when it is confidently wrong, and when the right move is to not use it at all. That is a skill, not an attitude, and skills have to be taught, practised, and corrected, the same as any other.

💡 Practical Takeaway

Pick one assignment this term where students must show their AI interaction alongside their final answer, prompt, output, and what they changed and why. That single artefact tells you more about a student's actual judgment than any policy document will.

Why "banning it" and "ignoring it" fail the same way

Schools tend to land on one of two positions, ban AI outright or say nothing and let it happen quietly, and both fail for the same reason: neither one puts an adult in the room while the student is actually deciding whether to trust what the model produced. A ban does not stop the behaviour, it just moves it out of view. Silence does not protect students, it just means the first time anyone examines their AI judgment critically might be in a job interview, not a classroom.

Why this matters for Malaysian schools specifically

The 2027 AI curriculum mandate is going to force this conversation for every school regardless of preference, and the schools that get ahead of it now will be teaching judgment as a habit rather than scrambling to retrofit it as a compliance requirement. That distinction, habit versus compliance, is usually the difference between a programme that actually changes how students think and one that produces a folder of evidence for an inspection.

Want a second opinion on where your students actually stand?

I run short diagnostic conversations with school leadership teams: not a sales pitch, a look at what your students are actually doing with AI right now, and what a judgment-first programme would need to look like for your context.

Book a Discovery Call

What This Means for You

The tools will keep changing, they always do, and whatever specific model or app is popular this year will not be the one students are using in three. Judgment is the part that transfers. Schools that treat AI literacy as "which tool, which policy" will be rewriting their approach every year. Schools that treat it as "how do you know when to trust this" are teaching something that outlasts the tool entirely.