A few months ago, on the road out past Lake Eildon, the sky did what Victorian skies do without warning: it dropped seven degrees, the light went flat, and the bitumen turned that shade of dark that tells you, before you consciously register it, that the surface is about to change. No app had warned me. No checklist covered it. What mattered was reading the road, the wind, the barometric pressure, the way the environment felt around, and adjusting not to a rule, but to a specific moment. That’s situated judgement, and it’s almost entirely contextual.
We talk a lot, in education and elsewhere, about frameworks, competencies, and rubrics as if good decision-making were a matter of applying the right rule at the right time. But anyone who has had to decide something under pressure knows this isn’t quite true. The rule is only ever a starting point. What determines whether a decision is good is whether it fits the specific, often unrepeatable situation in front of you. Recognising your own experience and judgment is essential, as education transfer has its limits. The rest is learnt in thousands of different contexts, some new, some the same, but always inductive; you never really know until you enter.
This is where AI’s limits become interesting rather than disappointing. Large language models are extraordinary at context-free reasoning: grammar, statistics, the general, probable case. Ask an AI tool for advice on a delicate conversation, and it will often produce something technically sound and completely misjudged: correct in the abstract, wrong for the actual humans involved. It doesn’t know that the student’s silence in class means something different this week than it did last week. It hasn’t stood in the room or experienced human nuance.
That gap is precisely what judgement is. Not the possession of the right general answer, but the skill earned slowly, through trial and error, of knowing which answer applies here, now, to this person, on this road, in this weather. It’s why an experienced rider reads a corner differently from someone quoting the road rules, and why an experienced manager reads a meeting differently from someone quoting the policy manual. Your discernment is what makes decision-making meaningful, and it remains irreplaceable.
There’s a temptation, faced with AI’s fluency, to think judgement itself is becoming obsolete: that if a system can produce a plausible, convincing answer instantly, the human contribution shrinks to simply checking its work. I think the opposite is happening. As AI absorbs more of the general cases, your unique situated judgment becomes more visible, not less. It’s the residue that doesn’t compress into a pattern. It’s what you bring when you subtract the general case from the actual one, and that remains vital.
Riding a motorcycle teaches this daily, in small ways: the corner that looks like every other corner but isn’t, the gravel patch that wasn’t there last time. Context teaches it in slower, human ways: the meeting that looks like every other meeting but carries a different weight. AI will keep getting better at the general case. That’s fine; it should. But the work of noticing when this case isn’t the general one, and acting accordingly, stays stubbornly, usefully human and remains essential.
Leave a Reply