There is a tempting product fallacy in AI right now: the better the one-shot output, the better the experience. For adults summarising a contract or generating a landing page, that's often true. The magic is efficiency. The less friction, the better. But children are different. And making games is different.
When a child uses a vibe-coding tool, the first generation does need to feel magical. Type "make me a spooky platform game where a frog collects stars in space" and get something playable within seconds. That spark says: your idea can become real. But the game should not be so finished that the child has nowhere to go next.
"If the AI produces the perfect thing, the child becomes an audience member. If it produces something good, surprising, and slightly imperfect, the child becomes a creator."
Children learn by making, not by receiving
Jean Piaget argued that children don't simply absorb knowledge , they actively construct understanding through exploration. Seymour Papert pushed that further with constructionism: children learn especially powerfully when they're building public, shareable things. A program, a robot, a turtle that draws. A thing they can point to and say: I made this.
If we turn AI into a machine that completes the child's idea for them, we undo the deepest lesson of constructionist technology. The better pattern is for AI to help the child externalise an idea quickly, then make that idea available for tinkering. The AI gets them over the blank page. Then the learning begins.
The lesson from Scratch, micro:bit, and Mindstorms
The best kid-coding tools of the last decades didn't work because they produced perfect outputs. They worked because they created playful environments where children could make, test, break, repair, remix, and share. Mitchel Resnick's Scratch model is a creative learning spiral: imagine, create, play, share, reflect, then imagine again.
The loop is not PROMPT → GENERATE → DONE. The loop should be IMAGINE → GENERATE → PLAY → CHANGE → SHARE → REFLECT → GENERATE AGAIN.
Wizard, planner, or compass?
A wizard can help , it reduces fear and gives structure. But too much wizarding changes the texture of the experience. Instead of discovering, the child is complying. Instead of asking "what happens if?", they're answering a form. The ideal experience should feel less like a rigid wizard and more like a compass: enough direction to keep moving, enough feedback to avoid getting lost, enough agency to choose the next step.
The destination should surprise you
Vibe coding can give kids something as powerful as Logo's turtle or micro:bit's sensors: a playable idea that appears almost instantly, and then invites them to shape it. Not a machine that says "here is your finished game," but a companion that says "here is your game. It works. What should we change next?" That "what next?" is everything.