Learn with AI: a patient tutor that makes you do the thinking

Build a lesson around a question

Choose something specific: how aperture changes a photograph, how spreadsheet references work, or why a circuit needs a resistor. State what you know and where you get lost. Ask for a plain explanation followed by a question that reveals whether you understood it.

UNESCO's guidance emphasizes human agency and thoughtful educational design when using generative AI. A practical implication for this guide is to keep the learner actively explaining, attempting, and checking, rather than passively receiving polished answers. UNESCO: Guidance for Generative AI in Education and Research

A useful twenty-minute session

Spend the first few minutes answering a baseline question without help. Ask for a short explanation tailored to your mistake. Next, work through one example with hints. Then try a different example on your own. Finish by explaining the idea in ordinary language as if you were teaching a friend.

Teach me the difference between voltage, current, and resistance using a low-voltage example. Start by asking what I already understand. Ask one question at a time. If I am wrong, give a hint before the answer. Finish with a new problem and check my units.

For technical or safety-sensitive topics, use the assistant to understand a reputable reference, and follow the reference's actual safety instructions. An analogy can illuminate an idea while still being incomplete. Ask where the analogy stops working.

Make the assistant adapt

If the explanation is too dense, ask for shorter sentences and one new term at a time. If it is too simple, ask for the mathematical relationship or a counterexample. An adult learner does not need childish language; clear wording and respect for existing experience are enough.

For language practice, specify the setting and level. For history, ask which statements are established facts and which are interpretations. For coding, type or modify the code yourself and predict the result before running it. Different subjects require different kinds of evidence that learning occurred.

Avoid the illusion of mastery

Recognition is easier than recall. An explanation can feel obvious while you are reading it, then disappear when you try to use it. Close the answer and reproduce the method. Return later with a new example. Keep a small error log rather than saving every conversation.

If the assistant marks an answer wrong, ask for the reasoning and compare it with the reference. It can make grading mistakes too. Use its availability to increase practice, not to outsource the final definition of correctness. Choose the next lesson from a demonstrated gap rather than from whichever subject the assistant happens to propose.