How to use AI tools wisely and safely
AI tools (chatbots, translators, image generators) can help a lot with learning – but only if you know how to ask them, how to check the answer and what you must never send them.
This page describes the current situation (November 2026) and is updated once a year.
Before you start: age and rules
- In their terms of use, many AI services require the user to be at least 13 years old (some even 18), and for minors they often require parental consent. Before opening an account, read the terms together with a parent.
- Check the rules of your school and teachers: when AI is allowed (for example for explaining the material) and when it isn't (tests, competitions, homework you should do yourself).
- No account is needed for the exercises on this website.
How to ask a good question
The instruction you give an AI tool is called a prompt. The clearer it is, the better the answer. A good question has:
| Part | Example |
|---|---|
| who you are and what you know | “I'm in 6th grade and I know what fractions are.” |
| what exactly you are asking for | “Explain to me how to add fractions with different denominators.” |
| how you want the answer | “On one example, step by step, without the ready-made solution to my homework.” |
| a check | “At the end, give me one task to check whether I understood.” |
Bad: “fractions”. Better: “I'm in 6th grade. Explain to me on one example, step by step, how to add 1/3 and 1/4, then give me a similar task for practice.”
Check the answer: AI can “make things up”
AI tools sometimes confidently write something that isn't true – they make up a date, a person, a book or the result of a calculation. This is called a hallucination. So:
- Compare with a textbook or a reliable website.
- Check calculations yourself or with a calculator.
- Ask for a source – and check whether that source really exists.
- Try the code before you believe it works (more about this on the page AI in learning programming (SR)).
- If an answer sounds strange or “too good” – ask an adult.
Privacy: what you never write to an AI tool
- your full name, address, phone number, the name of your school and class
- passwords, PINs, card details
- photos of yourself and other people without their permission
- other people's messages and secrets
What you write can be stored on the company's servers. Act as if you were writing on a public notice board.
Learning with AI, not instead of you
| Good use | Bad use |
|---|---|
| “Explain this term to me in a different way.” | “Do my homework for me.” |
| “Give me 5 questions to check my knowledge.” | copying answers without reading them |
| “Where is the mistake in my solution? Don't give me the whole solution.” | handing in AI text as your own |
| “Help me make a study plan.” | believing every answer without checking |
The brain learns when it makes an effort. If AI does the work for you, you will get a solution – but you won't learn. Use it as a patient helper, not as someone who writes instead of you.
Fake images, videos and voices
Today AI makes images, video and voices that look and sound real (deepfake). Before you believe or share something:
- check who published it and whether reliable media write about it;
- look at the details: hands, fingers, text in the background, shadows, unnatural lip movements;
- ask yourself: does the content want to make me angry or scared so that I share it quickly?
- never make fake images or videos of other people – it can hurt them and can be a crime.
Activity: fix the question and find the mistake
- Fix the question “geography test” so that the AI's answer is useful for learning.
- An AI wrote: “Nikola Tesla was born in 1856 in Belgrade and invented the light bulb.” Find the mistakes.
Solution
1) For example: “I'm in 7th grade, studying for a geography test about the rivers of Serbia. Ask me 5 questions one at a time, and after each of my answers tell me whether it is correct and why.”
2) Tesla was indeed born in 1856, but in Smiljan (in today's Croatia), not in Belgrade; he didn't invent the light bulb – his best-known inventions are related to alternating current and the induction motor. So the answer mixes true and false – that is why you check it.
Show the same question asked badly and well, then compare the answers. Then students check, in groups, an AI answer prepared in advance with deliberate mistakes.
Common misconceptions
- “If AI says so, it must be true” – AI can make a mistake or make up a fact; always check.
- “AI doesn't remember what I write to it” – many services keep conversations; don't send personal data.
- “Using AI is always cheating” – it depends on the rules: explaining the material is fine, handing in someone else's (AI) work as your own isn't.
Practice tasks
Try it yourself first, and only then open the solution.
- Write a good prompt that asks an AI tool for help with a math task without giving you the ready-made solution.
- Name four pieces of information you should never write to an AI tool.
- What is an AI tool's “hallucination”, and how do you protect yourself from it?
Solution 1: example
“I'm in 8th grade. I have a task with the Pythagorean theorem and I don't know how to start. Don't give me the solution – ask me a question that will lead me to the first step, and wait for my answer.”
Solution 2: answer
For example: your full name and address, phone number, password, payment card details, the name of your school and class, photos of other people without permission.
Solution 3: answer
It is when AI confidently writes incorrect information (a made-up date, person, source). You protect yourself by checking: a textbook, reliable websites, your own calculation, trying out the code, asking a teacher.
What's next
You have finished the first module! In the next module, How a machine learns, you will see how a model is made from data – and you will make your first model without programming. Until then, also take a look at the page Artificial intelligence in learning programming (SR).
Check your knowledge
Open the quiz
Quiz: using AI wisely and safely
1. What is a “prompt”?
2. What is an AI tool's hallucination?
3. Which information may you write to an AI tool?
4. Which use of AI helps learning?
5. What do you do first when you see a shocking video of a famous person?