academic world .wordpress.com

BLOG NUMBER 415 * There was a time when being able to explain a difficult concept beautifully could make you one of the strongest teachers in the building.
That advantage is shrinking.
A student can now get an explanation of photosynthesis, quadratic equations, colonialism or Shakespeare in seconds. They can ask for another example, request simpler language, generate a diagram, translate it, hear an analogy and ask the system to explain it again without becoming impatient.
The teacher therefore needs something beyond explanation.
UNESCO now describes education as operating within a teacher, AI and student relationship, rather than the traditional teacher and student relationship alone. The OECD similarly warns that access to generative AI does not automatically produce stronger thinking. Overdependence can reduce independent problem solving when learners accept generated answers without interrogating them (UNESCO, 2024; OECD, 2025).
The next generation will need teachers who can do these six things. 1.Learn Error Archaeology: Find the Thinking Beneath the Wrong Answer.
A weak instructional response to a wrong answer is another explanation.
A stronger response is diagnosis.
Imagine a Primary 5 pupil writes:
3/4 + 2/5 = 5/9
You could immediately explain how to find a common denominator. But first ask:
“Show me what you did.”
Then:
“Why did you add the 4 and 5?”
The pupil may reveal something important: When adding numbers, I add the top numbers and the bottom numbers.
Now you know the problem is not carelessness. The pupil has constructed a rule.
This is what I call Error Archaeology. You excavate the reasoning that produced the visible mistake.
The skill becomes even more important in an AI classroom because students can obtain correct answers without exposing incorrect mental models. A student may submit an excellent paragraph and still be unable to explain the argument inside it.
Create an Error Log during lessons. Do not only record who got an answer wrong. Categorise the error.
Was it a vocabulary problem?
A missing prerequisite?
A false rule?
A procedure applied in the wrong context?
A question interpretation problem?
A correct idea executed badly?
Then adapt teaching accordingly.
For example,* after a science question, instead of saying, “No, that is incorrect,” ask three students who selected different answers to defend their reasoning. Your next teaching move should be determined by what their reasoning reveals.
EEF describes adaptive teaching as uncovering learning rather than merely covering content. Teachers need evidence from learners before deciding whether to reteach, remove support, extend the task or address a misconception.The future teacher will not ask only:
“Did they get it?”
The better question is:
“What kind of thinking produced what I am seeing?”
That is a much more valuable professional skill than repeating yesterday’s explanation louder 2 Learn Thinking Exposure: Stop Showing Only the Finished Answer.
Many teachers model products.
Very few model cognition.
You solve the equation correctly on the board. You write the beautiful introduction. You pronounce the word accurately. You produce the perfect diagram.
Students see the result.
They rarely see the decisions.
That distinction matters.
Suppose you are teaching essay writing to Year 9 students. Instead of presenting a polished introduction, place an unfamiliar question on the board and say:
“I am not writing yet.”
Then expose your thinking.
“What exactly is this question asking me to judge?”
“Which word changes the direction of the essay?”
“I have three possible arguments. I am rejecting this one because I cannot support it with enough evidence.”
“I am choosing this example, but notice why I am not using it first.”
Now the lesson has moved from watch my answer to watch my judgement.
stress teaching pupils to plan, monitor and evaluate learning, while research on dialogic teaching shows the value of classrooms where pupils reason, argue, discuss and explain rather than merely produce short responses .
Try a Decision Commentary once every lesson.
Take one task and verbalise five decisions an expert makes while completing it.
Then give students another task and ask them to annotate their decisions:
I noticed…
I considered…
I rejected… because…
I changed… because…
I checked… by…
This works in mathematics, writing, science experiments, comprehension, art, coding and even early primary classrooms with simpler language.
The next generation does not desperately need another person who can produce an answer.
Machines can increasingly do that.
They need adults who can reveal how knowledgeable humans decide what to do when the answer is not obvious.
That is a different profession. * 3.Learn Friction Design: Know When Not to Explain Yet.
One of the strangest skills future teachers will need is the discipline to delay help.
Good teachers naturally want to rescue learners.
A student hesitates.
We explain.
A group gets stuck.
We give a clue.
Someone asks, “Sir, what should I write?”
We tell them.
Eventually the classroom becomes efficient, but the teacher is doing much of the difficult thinking.
This is where Friction Design matters.
Friction is the carefully controlled difficulty that forces the learner to retrieve, compare, predict, test or reconsider before receiving help.
Consider a Secondary 2 science lesson on floating and sinking.
Do not begin with ten minutes explaining density.
Place four objects where everyone can see them. Ask pupils to rank them from most likely to least likely to float.
Make every learner commit to a prediction.
Then test the objects.
Now ask:
“Which result damaged your original rule?”
Suddenly the explanation has somewhere to land.
Use a simple sequence:
Predict. Commit. Test. Explain. Revise.
You can use the same architecture elsewhere.
In literature: predict a character’s next decision before reading.
In mathematics: estimate the answer before calculation.
In history: choose the most likely cause before examining evidence.
In phonics: attempt the unfamiliar word before receiving the correction.
This is not an argument against explicit teaching. Students still require expert instruction and appropriate scaffolding. The question is when explanation enters the learning sequence.
Research on generative AI makes this issue even more urgent.*reviews warn that instant AI assistance can improve task performance while creating risks for independent thinking and long term skill development when users become dependent on generated solutions (OECD, 2025).
Before explaining, therefore, ask:
“What intellectual work can students safely do first?”
Do not remove every obstacle.
Some obstacles are where the learning is hiding * 4.Learn Question Architecture: Ask Questions That Are Expensive to Outsource.
Teachers were trained to ask questions that test whether students know the answer.
The next challenge is designing questions for which possessing an answer is not enough.
Ask:
“What is photosynthesis?”
A search engine can answer it.
Ask:
“Write five characteristics of democracy.”
AI can produce them immediately.
Ask:
“Explain the causes of World War One.”
A student may submit a sophisticated response without owning a single sentence of the reasoning.
The answer is not to ban every tool.
Upgrade the question.
I call this Question Architecture.
Take an ordinary question and add one or more intellectual constraints.
Local constraint:
“Use two examples from our community.”
Evidence constraint:
“Which piece of evidence most strongly supports your conclusion?”
Trade off constraint:
“What would we gain and lose if we chose this solution?”
Counterexample constraint:
“Give one situation where your rule would fail.”
Decision constraint:
“You may choose only one option. Defend it.”
Revision constraint:
“Here is new evidence. What part of your original answer must now change?”
Imagine a business studies class.
Instead of:
The future belongs to teachers who can create questions more intellectually valuable than the answers machines can generate.. * 5. Learn Proof of Thinking: Stop Confusing Beautiful Work With Learning.
A polished assignment is becoming weaker evidence of learning.
That sentence will become increasingly important.
Imagine two students submit excellent essays.
Both are grammatically strong.
Both contain evidence.
Both have impressive vocabulary.
One student understands every decision inside the essay.
The other generated most of it and cannot reconstruct the argument.
If your assessment sees only the final product, both students may receive similar marks.
The teacher now needs Proof of Thinking.
After important tasks, introduce a short verification layer.
Ask the student:
“Which part of your answer was hardest to decide?”
“Show me one thing you changed.”
“Which evidence would you remove if I cut the word count by 30 per cent?”
“I disagree with your conclusion. Defend it.”
“Apply your method to this new example.”
This can take sixty seconds.
For larger assignments, collect three pieces of evidence:
The product: What did you finally create?
The process: What drafts, decisions, calculations or revisions produced it?
The defence: Can you explain, adapt or justify the work live?
A mathematics teacher might change one number after the student submits a solution.
A geography teacher might introduce new data.
An English teacher might ask for an alternative ending.
A science teacher might ask which variable would invalidate the experiment.
assessment highlights practical activities, oral examinations, portfolios and student defences as ways of capturing competencies that conventional written products may not fully reveal.
Do not become obsessed with catching AI use.
Design assessments that make thinking visible.
A student who genuinely owns the learning should be able to explain it, alter it, defend it or use it somewhere else.
That is stronger evidence than a beautiful submission. * 6.Learn Cognitive Handover: Your Students Must Eventually Need Less of You.
Here is an uncomfortable test of teaching quality:
What happens when you stop helping?
Some classrooms look excellent while the teacher is present because the teacher is carrying almost every cognitive responsibility.
The teacher identifies the problem.
The teacher chooses the strategy.
The teacher reminds students what comes next.
The teacher notices errors.
The teacher decides whether the work is good enough.
The teacher tells everyone how to improve.
Remove the teacher and performance collapses.
That is not yet independence.
The next generation needs teachers who understand Cognitive Handover.
Think of responsibility as something you deliberately transfer.
At the beginning of a new skill, you may model heavily.
Later, remove one support.
Instead of giving students the steps, ask them to reconstruct them.
Instead of identifying the mistake, say:
“One part of this solution conflicts with something we learned yesterday. Find it.”
Instead of telling them which strategy to use, ask:
“You have three strategies available. Which fits this problem best, and why?”
Eventually, students should be able to plan, monitor and evaluate their own work help pupils develop those processes and gradually become more independent users of learning strategies.
Use a weekly Support Audit.
Choose one repeated classroom task and ask:
“What am I still doing here that students could now do?”
Maybe students can set the success criteria.
Maybe they can select a method.
Maybe they can check their own first draft against an exemplar.
Maybe groups can decide when they need teacher intervention.
Explanation creates access.
But the final purpose of good teaching is not permanent dependence on the explainer.
Your classroom should gradually produce learners who can think when you are no longer standing beside them.
That may become one of the clearest distinctions between teachers who survive the next educational shift and teachers who lead it.