We teach children to think together with AI: set the task, check the answer and carry an idea through to their own project — a fairy tale, a cartoon, a bot or a piece of research. Every lesson ends with a finished piece of work.
Thinking First.
AI Second.
AI amplifies people who know how to think and doesn’t rescue those who don’t. So before any tools, we teach thinking: how to break a task down, what to hand to the machine and how to check the answer.
The tools will change within a year. The ability to set a task, spot a mistake and carry an idea through to a result will stay.
Writes a prompt that gets the intended result — and improves it when the result isn’t right
Sorts tasks into three groups: I do it myself · I do it with AI · I never give it to AI
Doesn’t take AI at its word and knows three ways to check: the textbook, sources, a second AI
Makes text, pictures, a song and a short video with AI — and puts them together into one story
Builds a personal “School AI Kit” with a clear line between learning and copying
Defends their own project: what they did, why, which robots helped and what they learned
The lesson topics are shared; the delivery is not. A six-year-old dictates out loud while a grown-up types; a teenager works alone and builds a real bot or a piece of research.
Games, cards and movement. The child speaks, a grown-up types. Project: a fairy tale with pictures and a song.
30 minutes · up to 6 kidsTemplates and examples, work by model. The child is at the keyboard, a grown-up nearby. Project: a cartoon or a comic.
45 minutes · up to 10 kidsCase reviews and independent work. Project: research, a bot, an automation or a media project.
60–75 minutes · up to 15 kidsShow a toddler an apple 20 times and on the 21st they’ll pick it out from other fruit. AI is the same: it learns from examples. Good examples give good results; bad ones give confident mistakes.
That’s how the first lesson starts. From then on a child neither fears AI nor trusts it blindly — they understand where its answers come from.
Every lesson ends with a finished piece of work, every block with a result you can show at home, and the course with the child’s own project.
We build learning games for the course, for schools, camps and company events. The child doesn’t sit through a lecture about AI — they play through levels and pick up a skill on the way.
You → text robot → picture robot → video robot. Project passport: topic, audience, steps, deadline.
The project is 80% done. We learn to fix typical snags: the style “drifts”, the AI made up a fact.
Pair reviews and a portfolio with all the work from the course.
A presentation in front of parents, certificates and a talk about what stays human in any job.
Once a week — a steady pace across a school term.
Twice a week — the same program, faster.
A holiday intensive camp: two lessons a day and a project at the end.
Together with the Leader Educational Center we ran AI activities for children and parents at the Science Festival organized by the National Academy of Sciences of Belarus.



A child takes a photo, picks a career and a few seconds later sees themselves in it — for example, as a space biologist in a spacesuit.
The picture is only the way into the conversation. Then we explain the main thing: what goes first is not jobs but tasks within them. So it matters to understand what a human does in the job and why a machine can’t do it.
A children’s program for the families of DHL Express Qatar employees: from a first meeting with artificial intelligence to their own project and a graduation at the company’s office.
A joint course for children aged 6–18: 12 lessons, 3 age tracks, a final project, a portfolio and a certificate.
lider.by/ai-vunderkindy-curs (in Russian)School and university students intern with us: they work alongside the team on real tasks, with a mentor, and leave with a real case for their portfolio rather than a classroom exercise.
Lesson 2 “The language of robots: prompts” and Lesson 4 “Checking: AI makes mistakes”, each as a short animated recap: what the lesson covers, what the kids do and what work it leaves behind.
Open any lesson: what happens in class for each age, the homework and how it connects to school.
Every lesson leaves a finished piece of work. By the end of the course they add up to a portfolio, with the child’s own project at its center.
Most AI services are designed for users aged 13 and up. So the youngest work only through the teacher’s screen or a learning account, and the rules are discussed in the first lesson and repeated in the fourth.
Quite the opposite — a whole block of the program is built on this. Children learn to sort tasks into three groups: I do it myself because I’m learning; I do it with AI because it speeds things up; I never hand it to AI. A separate lesson is about AI making mistakes and three ways to check its answer.
The rule is set in the very first lesson: tasks that test a child’s own knowledge are done without AI, and if AI helped, they say so openly. We teach children to use AI as a tutor who explains, not as someone who hands in the work for them.
Not if the format is right. At six, a child doesn’t type or sit alone with a service: they say out loud what they want, and a grown-up types it in. The lesson lasts 30 minutes, in a group of up to six, through games and cards. The main thing they learn is to put what they imagine into precise words.
No. It’s preparation for any career. A child who has learned to set a task and check the result will use that in medicine, design and business alike.
They certainly will. That’s why the program is built not around specific tools but around skills: setting a task, breaking it into steps, checking the answer and carrying an idea through to a result.
12 lessons · 3 age tracks · groups of up to 6, 10 and 15 kids · their own project, a portfolio and a certificate. Ready-made materials in English and Russian.
ina.nistoras@magnaqore.io · +375 44 735-23-52 · magnaqore.io