Using AI in your learning · A classroom guide for MYP and DP students

AI tools can genuinely support your learning. They can re-explain a tricky idea, help you prepare before an assessment or give you fast feedback on an idea.

This is not a list of bans; it is a map to help you navigate learning supported by and not inhibited by AI.

To use that map well, it helps to be clear about one thing first: what learning actually is, and why it cannot simply be handed to you.

Why the work matters - the ladder

Learning is something you do, not something you receive. And it asks more of you the further you go.

Knowing

Holding facts and information in mind.

Command terms

definestateidentifyoutline

Understanding

Grasping how those facts connect, and what they mean.

Command terms

explainanalysedistinguishsuggest

Synthesising

Combining them into something new, and using them in unfamiliar situations.

Command terms

evaluatediscussjustifyto what extent

more cognitive work, more durable learning

Climbing this ladder takes real effort, and that effort is not a side effect: it is where the learning happens. The research is strikingly consistent on this. Things stick better when you make studying a bit harder rather than easier¹: for example, by spreading practice out over time¹ and, above all, testing yourself instead of just rereading.¹ Understanding, in particular, cannot be handed over: it grows out of doing the harder thinking.

This is exactly why AI needs care. A chatbot can hand you an answer that already looks finished and polished, without you having done any of the climbing, and it becomes easy to mistake that polish for your own understanding.² Every hard task has a bit of struggle you cannot skip if you really want to learn. So the rule running through everything below is simple: use AI for the work around the climb, and do the climbing yourself.

Each situation pairs a short note on when and how to use AI with why it matters for your learning, and names the Ecolint Learning Principle behind it. Beneath each one are specific suggestions for the tools that fit: an LLM (a general chatbot) and NotebookLM, which answers only from sources you give it, such as your own notes and readings.

Sources1 Kuepper-Tetzel (2023)2 Furze (2026)

Generative AI Acceptable Use ScaleFull scale

0 No AI1 AI Planning2 AI Collaboration3 Targeted AI4 Full AI

Example prompts are written for

The advice on every card is the same whichever subject you pick; only the quoted example prompts change.

Before a lesson

Getting ready, so the lesson is not your first contact with the content.

Getting ready for a lesson

Level 3 · Targeted AI

When and how to use AI

Before a lesson on a new topic, use AI to get the lay of the land: a plain overview of the main syllabus points, the key terms you will meet, and the big questions the topic is trying to answer.

Why it matters for your learning

Walking in with a rough map means the lesson is not the first time you meet the ideas, so you can follow the harder parts and ask better questions instead of just trying to keep up. A little foundational knowledge in place is part of what lets you think at a higher level once you are in the room.

AI suggestions

Ask an LLM to preview the topic and provide the key terms and big questions in plain English.I have an IB DP Geography lesson on the first sub-unit of the Population unit. In plain English, give me the key terms I will meet and the three big questions this topic is trying to answer.I have an IB DP ESS lesson on the first sub-unit of the Foundation unit. In plain English, give me the key terms I will meet and the three big questions this topic is trying to answer.I have an MYP year 4 maths lesson introducing quadratic functions. In plain English, give me the key terms I will meet and the three big questions this topic is trying to answer.
Use NotebookLM to hold any slides or resources your teacher shares ahead of time, then ask it for a short overview and some questions to consider.

Generative AI Acceptable Use ScaleFull scale

Level 3 · Targeted AI

AI does one bounded job here, giving you a plain preview to read critically before the lesson rather than work you hand in.

Ecolint Learning PrinciplesOpen principles

Learning occurs at various levels of complexity.

Cognitive overload causes inefficient, ineffective learning.

During a lesson

The heart of it: using AI well in the ordinary, daily work of learning.

Answering questions in a lesson

Level 0 · No AI

When and how to use AI

Answer from your own head first. A question in class is there to find out what you already understand and to push your thinking a little further. If a single word genuinely blocks you, a quick AI definition or translation to get going is fine. The reasoning, the example, the link to the bigger topic: keep that yours.

Why it matters for your learning

Class is where you find out what you actually understand, by saying it out loud and seeing how it lands. Answering from your own memory is part of what fixes an idea in your mind.¹ Hand the question to a chatbot, and you get an answer, but you skip the bit of thinking that would have made it stick.

AI suggestions

Ask an LLM only if a single term blocks you, for a one-line definition, then answer the rest yourself.I am an IB DP geography student. Define population dividend in one sentence.I am an IB DP ESS student. Define net primary productivity in one sentence.I am an MYP year 5 maths student. Define the discriminant in one sentence.

Generative AI Acceptable Use ScaleFull scale

Level 0 · No AI

The question is there to show and stretch your own understanding in the moment, so the thinking has to stay yours.

Ecolint Learning PrinciplesOpen principles

Building understanding is enhanced through a culture of thinking.

The ability to follow thoughts carefully and control one's attention are prerequisites for learning.

Sources1 Kuepper-Tetzel (2023)

When you can't follow the teacher's explanation

Level 3 · Targeted AI

When and how to use AI

This is a good moment for AI. Ask it to explain the idea more simply, from a different angle, or with a worked example. Then check your new understanding against your notes and your teacher. Use it to get back into the lesson, not to replace it.

Why it matters for your learning

When an explanation is too much to take in all at once, it does not stick, and a simpler version, or the same idea from a different angle, can clear the jam. A chatbot is also a low-pressure place to admit you did not follow something: it can feel easier to ask than to put your hand up, and we learn better when asking does not feel risky. Get yourself back into the lesson, then check what you now understand with your teacher.

AI suggestions

Ask an LLM to re-explain the idea, telling it where you got lost and asking for one worked example.Explain the demographic transition model more simply, as if I am meeting it for the first time, with one worked example.Explain negative feedback loops more simply, as if I am meeting them for the first time, with one worked example.Explain completing the square more simply, as if I am meeting it for the first time, with one worked example.
Use NotebookLM if the idea comes from a set reading, so its explanation stays tied to the actual text rather than inventing details or introducing too many new contexts.

Generative AI Acceptable Use ScaleFull scale

Level 3 · Targeted AI

AI takes on the single task of re-explaining a point you are stuck on, which you then check against your notes and your teacher.

Ecolint Learning PrinciplesOpen principles

Cognitive overload causes inefficient, ineffective learning.

Affectively and socially healthy environments promote learning.

Working with a challenging text

Level 1 · AI Planning

When and how to use AI

Read it yourself first, slowly, even when it fights you. If you are stuck on a word or a reference, look it up. Save AI summaries for after you have formed your own first impression; never use them instead of the reading.

Why it matters for your learning

A summary tells you what a text says; reading it yourself is how you come to understand it. The effort of working through writing that does not give up its meaning straight away is exactly where understanding gets built. Skip the reading and you end up with an opinion about a text you have never really met.

AI suggestions

Ask an LLM to look up a single unfamiliar term or reference, then get back to the reading.
Use NotebookLM after your own first read: add the text and ask it to clarify a confusing passage, with answers drawn from the source.Using only the text I have added, explain what this paragraph is arguing.

Generative AI Acceptable Use ScaleFull scale

Level 1 · AI Planning

You do the reading and form your own view first, using AI only to clarify a word or passage afterwards, never to stand in for the text.

Ecolint Learning PrinciplesOpen principles

Solving meaningful problems and transferring knowledge across domains contributes to deep learning.

The ability to follow thoughts carefully and control one's attention are prerequisites for learning.

Brainstorming at the start of a task

Level 1 · AI Planning

When and how to use AI

Get your own ideas down first, then bring AI in to widen the field, challenge your angle, or point out things you missed. Treat what it gives you as raw material to argue with, and pick your own direction. Keep a note of which ideas were yours.

Why it matters for your learning

Your first ideas matter because they come from your own experience and the links only you would make. AI is a useful thing to argue with once you have something of your own to test, but if it hands you the whole starting point, the task stops being yours.

AI suggestions

Ask an LLM once your own ideas are down, to widen the list or push back on your angle.I am planning an answer on why some pronatalist schemes succeed, and others fail. My three ideas are [list yours]. What angles am I missing?I am planning an answer on why some protected areas succeed, and others fail. My three ideas are [list yours]. What angles am I missing?I am starting a criterion B investigation into how the number of diagonals in a polygon grows as you add sides. My three ideas for finding the pattern are [list yours]. What have I not thought of?

Generative AI Acceptable Use ScaleFull scale

Level 1 · AI Planning

Your own ideas and direction come first, and AI only widens the planning, so nothing AI-made becomes the work itself.

Ecolint Learning PrinciplesOpen principles

All learning is personal.

Building understanding is enhanced through a culture of thinking.

Researching a topic

Level 1 · AI Planning

When and how to use AI

Use AI to get your bearings, suggest search terms, and explain unfamiliar ideas. Then go to reliable sources and check any fact, quotation, or figure before you use it. Never repeat a data point you have not checked.

Why it matters for your learning

AI tools sound confident even when they are wrong, and they can quietly pass on the biases in what they were trained on.¹ Checking, comparing and judging sources is part of the very skill you are at school to build. Let the tool point you towards good sources; do not let it stand in for reading them.

AI suggestions

Ask an LLM to get oriented and suggest search terms, then consult real sources and verify every claim.
Use NotebookLM to gather your actual sources, articles, case studies and data, so its answers come from your own material, which you still check.

Generative AI Acceptable Use ScaleFull scale

Level 1 · AI Planning

AI helps you get oriented and find search terms, but you go to real sources and verify everything, keeping the research yours.

Ecolint Learning PrincipleOpen principles

Solving meaningful problems and transferring knowledge across domains contributes to deep learning.

Sources1 Villar Onrubia et al. (2025)

Planning an essay

Level 1 · AI Planning

When and how to use AI

Start by mapping what the essay question demands. Then build your argument: your point, your main reasons, your evidence. Once that plan is down, use AI to test the structure or find weak links. The outline you actually write from should be your own.

Why it matters for your learning

Forcing an argument into shape is how you work out what you actually think, and that kind of problem-solving is where deep learning happens. A structure handed to you skips the part where the topic becomes yours.

AI suggestions

Ask an LLM once your own plan is mapped, to stress-test the structure rather than build it for you.Here is my plan for: Evaluate the extent to which integrated drainage basin management plans balance the needs of different stakeholders. Where is the argument weakest, and what have I left out?Here is my plan for: Evaluate the effectiveness of international agreements in reducing greenhouse gas emissions. Where is the argument weakest, and what have I left out?Here is my plan for my criterion D report on whether one large pizza is better value than two small ones. Where is the reasoning weakest, and what have I left out?

Generative AI Acceptable Use ScaleFull scale

Level 1 · AI Planning

You build the argument and outline yourself and bring AI in only to stress-test a plan you already own.

Ecolint Learning PrinciplesOpen principles

Solving meaningful problems and transferring knowledge across domains contributes to deep learning.

All learning is personal.

Understanding a concept through examples

Level 3 · Targeted AI

When and how to use AI

Ask AI for analogies, real-world examples, or the same idea explained three different ways. This is a strong, sensible use. Then test whether you can explain it back without looking, and try it on a fresh situation.

Why it matters for your learning

A good example or analogy lets you meet a tricky idea at a manageable level first, then build on it. Understanding tends to grow in steps, and it grows fastest when the next step is just beyond what you can already do, rather than far past it.

AI suggestions

Ask an LLM for analogies or the same idea explained a few different ways, then test yourself on it.Give me three different analogies for the multiplier effect in a growing city.Give me three different analogies for trophic efficiency in a food chain.Give me three different analogies for what a function does to a number, with an everyday example of each.
Use NotebookLM to build a set of worked examples from your own notes and have it quiz you on them.

Generative AI Acceptable Use ScaleFull scale

Level 3 · Targeted AI

AI's defined job is to generate analogies and examples, which you then test by explaining the idea back without looking.

Ecolint Learning PrinciplesOpen principles

Learning occurs at various levels of complexity.

Learning occurs in a space where objectives are just beyond but not too far from the learner.

After the lesson

On your own afterwards, where most of the temptation and consolidation live.

Hitting a wall and getting unstuck

Level 1 · AI Planning

When and how to use AI

When you are properly stuck, stay there a little longer than feels comfortable before reaching for help. Then ask the AI for a hint, a question in return, or the next small step, not the finished answer.

Why it matters for your learning

Every hard task has a bit of struggle you cannot skip if you actually want to learn it.¹ In fact, wrestling with a problem before anyone shows you how tends to leave you better prepared to learn it.¹

AI suggestions

Ask an LLM for a hint or the next small step, not the finished answer.I am stuck on this question about river discharge and flooding. Give me a hint to get moving, not the full answer.I am stuck on this question about ecological footprints and carrying capacity. Give me a hint to get moving, not the full answer.I am stuck on this question about solving simultaneous equations. Give me a hint to get moving, not the full answer.

Generative AI Acceptable Use ScaleFull scale

Level 1 · AI Planning

You sit with the struggle and ask AI only for the smallest hint to get moving, never the solution, so the solving stays yours.

Ecolint Learning PrincipleOpen principles

Making mistakes is a normal, inevitable and even fertile part of learning.

Sources1 Furze (2026)

Producing active recall cards to review

Level 3 · Targeted AI

When and how to use AI

Turn your notes or a text into active recall question-and-answer cards you can test yourself with, rather than a summary you reread. AI is genuinely useful for drafting these quickly. Check every card against your own notes: the questions and answers should come from what you have produced, not introduce new content.

Why it matters for your learning

Picking out what matters and putting it into your own questions is the real thinking, the part that turns a page of notes into understanding. Cards work because they make you pull the answer out of your own head rather than recognise it on the page,¹ and that effort is the point.

AI suggestions

Use NotebookLM to collect your lesson notes and resources, then turn your own materials into active-recall question-and-answer cards.From the notes I have added on the extreme environments unit, make fifteen active recall cards I can test myself with.From the notes I have added on the carbon cycle, make fifteen active recall cards I can test myself with.From the notes I have added on right-angled trigonometry, make fifteen active recall cards I can test myself with.
Ask an LLM to draft cards from a topic quickly, which you then check against your notes.

Generative AI Acceptable Use ScaleFull scale

Level 3 · Targeted AI

AI drafts the cards from your own notes as a set task, while you decide what matters and answer them from memory.

Ecolint Learning PrinciplesOpen principles

The ability to follow thoughts carefully and control one's attention are prerequisites for learning.

Solving meaningful problems and transferring knowledge across domains contributes to deep learning.

Sources1 Kuepper-Tetzel (2023)

Revising for a test

Level 4 · Full AI

When and how to use AI

Lean on AI here. Get it to make practice questions, build flashcards, set you a spaced schedule, or quiz you on last week's work, then close the screen and answer from memory. The one thing to avoid is reading its summaries instead of recalling the content yourself.

Why it matters for your learning

Pulling answers out of your own head is the hard part, and the hard part is what makes them stick. Testing yourself, rather than just rereading, makes things stick far better,¹ and spreading that practice over several days helps even more.¹

AI suggestions

Use NotebookLM to keep all your notes for a unit in one place, then ask for practice questions grounded in your own material.Using the notes I have added on urban environments, write ten exam-style practice questions and a short study guide.Using the notes I have added on biodiversity and conservation, write ten exam-style practice questions and a short study guide.Using the notes I have added on probability, write ten criterion A style practice questions and a short study guide.
Ask an LLM to quiz you out loud and mark your answers, then close the screen and recall from memory.

Generative AI Acceptable Use ScaleFull scale

Level 4 · Full AI

Revision is private practice, so you can lean on AI throughout to generate questions and quiz you, as long as you do the remembering yourself.

Ecolint Learning PrincipleOpen principles

The ability to follow thoughts carefully and control one's attention are prerequisites for learning.

Sources1 Kuepper-Tetzel (2023)

Writing and redrafting your work

Level 2 · AI Collaboration

When and how to use AI

Write the first draft yourself, in your own words, however rough it feels. Then use AI for feedback on the draft you already have: ask it to point out unclear sentences, gaps in the argument, or places where the evidence is thin. Decide which suggestions deserve acting on, make the changes yourself, and keep a note of how AI was involved.

Why it matters for your learning

Writing is not what happens after the thinking; it is the thinking. Sentences you did not write carry understanding you do not have, and a polished paragraph is easy to mistake for your own understanding.¹ Feedback on your own draft is different: it shows you how your writing lands with a reader, and every change you choose to make is a decision you have reasoned through.

AI suggestions

Ask an LLM for feedback once your draft exists, asking for weaknesses to fix rather than rewritten text.Here is my draft paragraph on flood management trade-offs. Do not rewrite it. Point out the two weakest parts of the argument and one place where the evidence is thin.Here is my draft paragraph on the trade-offs of building large dams. Do not rewrite it. Point out the two weakest parts of the argument and one place where the evidence is thin.Here is my written method for this criterion C question. Do not rewrite it. Point out the two steps a reader could not follow, and one place where I have not said why I did what I did.
Use NotebookLM to check a draft against your sources: add your notes and readings, then ask whether your claims match what the sources actually say.

Generative AI Acceptable Use ScaleFull scale

Level 2 · AI Collaboration

AI generates no new material; it responds to a draft you wrote, and you evaluate its feedback and make the changes yourself, acknowledging how you used it.

Ecolint Learning PrinciplesOpen principles

All learning is personal.

Making mistakes is a normal, inevitable and even fertile part of learning.

Sources1 Furze (2026)

Generative AI Acceptable Use Scale

Every situation above is mapped to a level on this scale. It shows, at a glance, how much AI involvement is appropriate for a given kind of work, and what you would need to acknowledge.

Adapted from the AI Assessment Scale (AIAS) by Furze, Perkins, Roe and MacVaugh (2024).

Level	What it means	What you acknowledge
0No AI	The task is completed without AI assistance. AI must not be used at any point. Demonstrate your own knowledge, understanding and skills.	No acknowledgement required.
1AI Planning	No AI content is allowed in the final submission. Use AI to brainstorm, structure and research; the final work shows how you developed and refined your own ideas.	A short statement outlining how you used AI.
2AI Collaboration	No new material may be generated by AI; AI is used to evolve your own content. Use it for idea generation, drafting support, feedback and refinement, then critically evaluate and modify its output.	A short statement outlining how you used AI.
3Targeted AI	AI may be used to complete a specific task or tasks set by your teacher. Engage with the AI content, evaluate it critically and provide human oversight.	Cite all AI-created content.
4Full AI	You may use AI throughout the task to support your work while demonstrating your own critical thinking. Treat AI as an assistant to your creativity, and oversee and evaluate everything it produces.	Cite your use of AI.

The scale in full: Ecolint source document.