The answer is no longer the whole achievement
Imagine two students submit equally polished explanations. One can identify the central assumption, show the evidence and explain what would change their mind. The other cannot explain why the conclusion follows. If we look only at the finished text, we miss the difference that education ought to care about.
OpenSkool's argument is that AI makes this distinction more urgent. The ability to obtain an answer and the ability to justify it should not be treated as the same achievement. Critical thinking becomes more valuable when plausible answers are easy to produce.
Some difficulty is part of learning
It is tempting to describe every reduction in effort as progress. But effort has different jobs. Reformatting a document may be an obstacle worth removing. Struggling to distinguish two explanations may be the learning itself. A tool can remove either kind, and the interface will not necessarily tell us which it has removed.
This is not an argument for making learning needlessly hard. A clearer explanation, translation or example can open a subject to someone previously excluded from it. The question is whether the assistance helps a learner participate in the thinking, or quietly replaces their participation.
Human agency needs something concrete to mean
UNESCO's guidance and student competency framework emphasise human agency and responsible engagement with AI. Our interpretation is that agency must show up in the work: choosing a question, challenging a premise, checking a source and taking responsibility for a conclusion. A button labelled ‘accept’ is not enough.
A learner should be able to disagree with the tool for a reason. They should also be able to discover that their own first answer was wrong. Critical thinking is not automatic suspicion of AI or automatic confidence in oneself; it is willingness to make both answerable to evidence.
Ask for the decision trail
One practical implication is to assess more than the final product. Ask students to retain an initial explanation, an AI-assisted revision and a short account of what they accepted or rejected. Invite them to identify a claim they could not verify. Change one assumption and ask whether the conclusion still holds.
These are editorial proposals, not claims of proven learning gains. They will need adaptation to age, subject, access and assessment rules. Their purpose is to make judgement visible without turning every lesson into a surveillance exercise.
- Which part of this answer is evidence, and which part is interpretation?
- What important alternative does it leave out?
- What observation would make you revise your conclusion?
- Could you explain the idea without reopening the tool?
But surely we have always used tools?
Yes. Books, calculators, search engines and other people already extend what we can do. The useful question is not whether assistance is legitimate in the abstract. It is what competence we want the learner to develop, and how the chosen assistance supports or obscures it.
Sometimes a complete worked example is exactly what a beginner needs. Sometimes producing the explanation is the task. An educational decision should start from that purpose, not from a blanket enthusiasm for the tool or a blanket ban.
A better question for the next lesson
Instead of asking only ‘Did the student use AI?’, ask ‘What can the student now understand, question and do?’ Then design a small opportunity to demonstrate it. The goal is not to preserve old assignments unchanged. It is to preserve—and improve—the intellectual work those assignments were supposed to make possible.
Start with a concrete habit: verify one consequential claim, explain a disagreement, or share a lesson with its limits. A learning platform should help people practise those acts, not merely supply more answers to consume.
Sources & further reading
- Guidance for generative AI in education and research
UNESCO. Guidance centring human agency and educational purpose. OpenSkool's proposed exercises are editorial suggestions, not validated interventions.
- AI competency framework for students
UNESCO. An educational framework covering human agency, ethics, techniques and system design; not a forecast of specific jobs.
Examples and exercises are illustrative unless attributed to a source. No independent expert review is claimed.
A correction, a counterexample or an experience worth sharing?
Join the conversation ↗