Online · kids and teens
Knowing what they really want. Sticking with it when it doesn't come out the first time. Staying the author with the AI at their elbow.
A lab is a few weeks on one project, in a group of no more than 8. The craft is different in each; the way we work is the same. Pick by age, and by what your child is drawn to.
Sessions are held in Russian, and so is the group chat between them. A child who speaks and reads it with ease will be fine, whichever country you live in. One who still has to work at it will get lost. Better to wait. You can write to us in English, and we'll answer the same way.
10+ · 10 sessions · Saturdays or Fridays
We take other people's games apart, invent our own, build a prototype with the AI and play it with the others.
Led by Yuri Yamshchikov, designer and producer of more than 100 published board games.
Ages 8–12 · 6 sessions
We dream up a hero and a story, find our own visual style, illustrate it and make a real book.
Led by Ksenia Leonova, journalist, screenwriter, teacher of dramatic writing.
Two more labs are in the works. Ask and we'll tell you →
One session a week. The work goes on the whole time.
Between sessions the kids can keep working on their project, with the AI and the lab leads close by. The finished thing grows out of that loop.
An hour live. We take apart a new technique, see what everyone came up with this week, try it together. Everything that matters happens here.
Between sessions the whole group lives in one chat, the AI included. Everyone sees who's inventing what, peeks in and picks up ideas, and the lab leads watch and help.
We read each child's exchanges with the AI and show them where they agreed with the first answer straight away, where they argued and checked, and where a decision is plainly their own.
Once a week we write to you about the session and your child's part in it. The craft, the AI, how the work itself went.
And what stays the same from lab to lab
The AI helps run through options, gather material, see the weak spot. The idea, the choices and the last word stay with the child, and that's the main thing we want them to see.
Few enough for the lab lead to get to each child's work, every session.
We practise noticing what's going on inside while we work: where it lit up, where it got tiring, where the urge is to drop everything. Then we lean on those signals to choose the next step.
The platform surprised me: room to work together and alone, plus support and dialogue outside the sessions.
My son remembers the enjoyment: the process, the group, talking with the people behind it.
Now I'm sure he can have a real dialogue with the AI, where he doesn't stop at the first answer and keeps working on his own idea.
When I heard about Maria and Michael's experiment, I wanted my son to try working with an AI under the eye of thoughtful adults.
Vlad got hooked on the back-and-forth, and the AI turned out to be a great way for him to open up, and for me to see inside. He never told me much himself about what he felt or imagined…
He started making up stories. I could see the thrill it gave him to watch his story come alive. And that made him want to write more.
Knowing that another kid was inventing something with the AI too made him even keener to try things himself.
It matters enormously to a child to be heard and included.
I hear my son discussing his project with enthusiasm, and he sounds very grown-up.
He feels needed. You can see it.
We're Maria and Michael, Lev's parents. It started with Lev and his friends: we wanted to show them how much they could make with their own hands. Then the AI came along, and a lot more became possible.
We began by showing them that a well-put request gets the AI to do what's needed. That turned out to be the easy part. The rest is harder. Does the child know what they actually want to create? Do they understand what it takes to get there? Do they see their own voice in what comes out?
Every lab runs on the same frame: the chat where everything happens, going over the AI conversations, the weekly letter to parents. That part is ours.
The labs are led by makers we'd gladly learn from ourselves, and who see things the way we do.
Who leads the labs
Yes, all of it: the hour on Zoom and the chat in between. A child who speaks and reads Russian without thinking about it will be at home here, whether you live in Riga, Lisbon or Moscow. One who still translates in their head will fall behind; for them it's too soon. Parents are another matter. Write to us in English.
We teach them to argue with it, check it, and keep only the words and ideas that are theirs. Every session we go over how the conversation went: where they took its suggestion straight away, and where they turned it over, understood it and made it their own. Nobody teaches this at school, and we think it's the skill that matters most when working with AI.
We run on Claude, made by Anthropic, and on ChatGPT from OpenAI. Nothing the kids write or say goes off to train somebody else's models. It stays inside the lab, so the AI remembers what each of them is working on and the help fits the person.
Yes, and we're glad when they do. The labs don't repeat each other. Each has its own craft: dramatic writing, illustration, game design. And the habit of finishing what they started and holding their own next to the AI only gets stronger from lab to lab.
No more than 8, so the lab lead can get to everyone's work.
Help them get connected the first time. After that, sit in on a session only if you're curious yourself.
Nothing that gets marked. But every session leaves a few things for them to try on their own, and whoever does something between sessions hears back from us. The ones who try end up noticeably further along by the finish.
There's a recording and a transcript. Watching it back is less fun than being there, but it's enough to catch up.
Get in touch. We'll write to you on the messenger you prefer, ask about your child and discuss what to try and what to expect.
Get in touch
