When AI is part of learning: making student thinking visible

Do you like this content? Get the interactive activity as a PDF to use in your classroom!

Imagine two students submit almost identical assignments. Both are well written. Both answer the question correctly. Both include thoughtful examples. At first sign, they both seem good.

But there is one important difference.
One student has gradually developed their ideas, questioned their first assumptions, revised their work, and learned through the process. The other asked an AI chatbot to generate a response, made a few small changes, and gave it to you, the teacher.

The final work may look very similar, but we know very well that the learning behind them is not. This is one of the challenges many teachers and educators are now facing.

Students have easy access to tools that can explain concepts, suggest ideas, summarise information, and answer questions in seconds. This may change some classroom routines, but it does not change how people learn.

Learning still happens when students ask questions, make connections, test ideas, explain their reasoning, receive feedback and, sometimes, realise that they need to rethink what they thought they understood. Learning is not only about arriving at an answer. It is also about what happens along the way.

This is why making students’ thinking visible matters more than ever. The aim is not to catch students using AI or to remove it from the classroom. It is to create learning experiences in which thinking, reflection and decision-making remain visible when AI is part of the process.

Why the thinking matters more than the answers

Teachers have always used students’ work to understand what they know and how they think. Today, this is more difficult. An essay, a summary, or a presentation may no longer show how a student arrived at their ideas. AI can help produce polished work very quickly, which means that the final product may tell us less about the learning behind it.

This does not mean that AI has no place in education.
It can support brainstorming, introduce different perspectives, help students improve a piece of writing, or provide feedback while they are working.

The most useful question is therefore not simply: “Did AI help with this task?” but probably: “What evidence of the student’s reasoning can I still see?”
The last question changes the focus: instead of looking only at the final answer, teachers and educators can pay attention to the decisions students made along the way. Why did they choose this example? Which ideas did they reject? Where did they change their mind? How did AI affect the direction of their work? What might they do differently next time?
These questions bring the learning process back into view.

What does the thinking process actually look like?

Making thinking visible does not mean asking students to explain every step of every task. It means noticing the signs that learning is happening. These signs often appear in small moments shared together out loud in the classroom:

A student explains why they chose one idea over another.
A student realises that an answer does not quite work and decides to try again.
Two students compare different viewpoints before reaching a conclusion.
A student asks a new question because something is not yet clear.
A group changes its mind after discovering new information during a discussion with classmates or the teacher.

These moments may seem ordinary, but they can tell us many things. AI does not remove these opportunities, but makes them harder to notice. Teachers and educators need to create moments in which students can talk about the questions, choices, doubts, and revisions that shaped their final response.

Try this in your classroom

The next time students complete a task with AI support, do not ask only for the final answer. Try asking:

What was your first idea?
Did AI suggest anything you had not considered?
Did you agree with everything it suggested? Why or why not?
What did you decide to keep, change or leave out?
Was there a point when you became unsure or changed direction?

These questions invite students to pause and think about their own learning, rather than simply describe what they produced.
This kind of reflection is often described as metacognition: thinking about how we think and learn.

Metacognition includes recognising what we already understand, noticing where we are confused, choosing a strategy, checking whether it is working, and deciding what to do next.

A student might realise:
“The AI explanation sounded clear, but I could not explain it without looking at the screen. I needed to go back and check whether I had really understood it.”
Another might say:
“AI suggested an argument I had not considered. I decided not to use it because I could not find enough evidence to support it.”

These reflections give the teacher useful information. They also help students become more active and independent learners.
The aim is not to ask students to report every thought they have. It is to help them notice the moments when they make a choice, encounter a difficulty, check their understanding, or change their approach.

Designing learning that makes thinking visible

Using AI in the classroom does not always require completely new activities. Often, a small change to a familiar task is enough to make students’ thinking easier to see.

You can ask: “Write an essay about climate change.”	Or you can try: “Ask AI to suggest three possible arguments. Which one do you find most convincing? Which one would you reject? Explain what influenced your choices.”
You can ask: “Summarise this article.”	Or you can try: “Write your own short summary, then compare it with one generated by AI. What did each version include or leave out? Create a final summary using the strongest parts of both and explain one change you made.”
You can ask: “Find the correct answer.”	Or you can try: “Compare two possible answers. Which one do you trust more? What evidence helped you decide? Is there one detail you would still check?”
You can ask: “Complete the task.”	Or you can try: “At the end of the task, identify one moment when you changed your approach. What made you reconsider your first idea?”

These small changes help reveal how students reached an answer, rather than showing only what they produced. They can also make the use of AI more purposeful. Students are not simply asking a tool to complete a task for them, they are comparing, checking, selecting, revising and explaining.

📷 Classroom snapshot 📷

A student asks an AI to explain photosynthesis. The explanation is accurate, clear and easy to understand. Rather than asking the student to repeat what AI has written, the teacher asks:

Which part helped you understand the topic better?
Was there anything you still did not fully understand?
Is there one statement you would like to check in another source?
How would you explain photosynthesis to a younger student in your own words?

The student now moves from reading an explanation to checking whether they have actually understood it. They may discover that they understand the general idea but are still confused about the role of sunlight, or may be able to repeat a definition but struggle to give an example. They may also notice that the AI explanation left out something discussed in class.

The goal is not to find out whether the student used AI.
The goal is to understand what they learned, what remains unclear and what they need to do next.

Interactive classroom activity: Follow the thinking

Time: 20–30 minutes

Age: Adaptable for different age groups

Group size: Pairs or small groups

Learning goal: Help students identify the choices, checks and revisions that make an AI-supported response a learning process.

Before the activity

Choose a short task linked to your subject. For example:

Explain why a historical event happened.
Solve a mathematical problem and explain the method.
Interpret the behaviour of a character in a story.
Suggest possible solutions to an environmental problem.

Prepare a simple “thinking trail” showing how a fictional student completed the task. For example, you could use a question about the French Revolution. The trail should contain five or six short cards or statements.

1. First idea	“I think the main cause was economic inequality.”
2. Question asked to AI	“What were the main causes of the French Revolution?”
3. AI suggestion	“The causes included social inequality, financial crisis, food shortages and Enlightenment ideas.”
4. Student reaction	“I had not considered food shortages. I want to check how important they were.”
5. Check or discussion	“The textbook shows that bread prices rose sharply before the Revolution.”
6. Final decision	“I will argue that economic problems became more serious because of social inequality and food shortages.”

Print the cards, display them on screen or write them on separate pieces of paper.

Step 1: Reconstruct the trail

Give each pair or group the cards in a mixed order. Ask students to arrange them in the order in which they think the learning process took place. They should be ready to explain:

Why did you choose this order?
Which card shows the student’s first understanding?
Where does the student’s thinking begin to change?
Which step connects the new information to the final answer?

There may be more than one reasonable order. The purpose is not to reconstruct the teacher’s exact sequence, but to discuss how ideas develop.

Step 2: Find the evidence of thinking

Once the cards are arranged, ask: Where can you see evidence that the student is thinking, rather than simply copying an answer? Invite students to point to specific moments, for example:

The student notices something they had not considered.
They decide that a claim needs to be checked.
They use another source.
They connect two different ideas.
They reject, adapt or expand part of the AI response.
Their final answer is different from both their first idea and the original AI suggestion.

Encourage students to support their answer with evidence from the cards:
“We think the learning becomes visible here because the student does not immediately accept the AI response. They decide to check one part of it.”
This is the central question of the activity. Students are not trying to identify whether AI was involved, they already know that it was - they are looking for the moments in which the learner made an active decision.

Step 3: Identify the missing step

Before giving the cards to the groups, remove one from the thinking trail. For example, remove the card in which the student checks the textbook and ask:
What seems to be missing?
Can the student move directly from the AI response to the final answer?
What could they do before deciding what to believe or include?
Which additional step would help us understand their thinking?

Groups then write their own missing card. They might suggest:

checking another source;
asking a follow-up question;
discussing the idea with a classmate;
asking the teacher for clarification;
explaining the idea without looking at the AI response.

Compare the cards created by the different groups. This helps students see that there is no single fixed learning process. Different learners may choose different strategies, but they should be able to explain why they chose them.

Step 4: Create a new thinking trail

Give each group a new classroom question. Ask them to create a short trail showing how a student could use AI without handing over the whole thinking process. Their trail should include:

a first idea or question;
one use of AI;
a moment of doubt, comparison or checking;
a decision made by the student;
a final response;
a short reflection on what changed.

Groups can exchange their trails and look for evidence of thinking in one another’s work.

Reflect together

End the activity with a short discussion:
Which step would be easiest to leave out?
What might happen if the student accepted the first AI response immediately?
Which strategies helped the student stay in control of the task?
How could you show a similar thinking trail in your own work?

The purpose of Follow the thinking is to help students recognise that learning involves choices: asking, checking, comparing, discussing, revising and deciding.

When these choices become visible, teachers can better understand what students are learning. At the same time, students can become more aware of how they learn and of the role they want AI to play in that process. When students learn to trace their own steps and notice where they got confused, or where they double-checked a claim, they move from passive consumers of technology to confident, independent thinkers. And as teachers, that's where the real work begins.

Further information and resources

Did you like this content? Get the interactive classroom activity as a PDF to use in your classroom!

Interested in learning more about AI? The Teacher corner has more to offer on the topic...

AI in the classroom

Generative AI

Deepfakes

Understanding mis- and disinformation