Young Builders Track (Ages 14-18)
Stop just playing the games — learn how the AI behind them works, and build your own.
Teenagers are already the heaviest users of AI in the country and almost none of them have been taught anything about it. They use it to finish homework, and school responds by banning it — which teaches nothing except how not to get caught. This track takes the opposite position: the interesting side of AI is not consuming it, it is building with it, and a fourteen-year-old who can build is in a completely different relationship to the technology than one who can only prompt. So we start from what they already care about — the recommendation feed, the game that adapts, the filter that finds a face — and open it up. Why does the model get it wrong? Where did the data come from? Who decided that? The projects are real and shipped, not simulations, because nothing kills interest faster than a fake exercise, and nothing builds it faster than showing a friend something you made that works. Mentors work with small groups: at this age the deciding factor is almost never ability, it is whether someone credible took you seriously. There is no prior coding requirement. There is a requirement to actually build something.
Prerequisites
- Ages 14 to 18
- No prior coding or AI experience required
- Curiosity and willingness to build, not just watch
- Parent or guardian registration
What you can do afterwards
- Explain in plain words how an AI model learns and why it gets things wrong
- Build and ship a working AI project of your own
- Use AI as a creative instrument in art, music or storytelling
- Spot a generated image, a biased answer and a confident falsehood
- Understand what happens to the data you give an app
- Present your own project to an audience and defend your choices
Sessions
Opening the black box
We start with the AI you already use every day — the feed that knows what you will watch next, the filter that finds your face — and take it apart. No formulas: the goal is that you can explain to someone at home how a model learns from examples, and why that explains both the magic and the mistakes.
Covered
Your first thing that actually works
By the end of this session you have built something and shown it to someone. It will be small. It will also be yours, running, and shareable — which is the moment most people stop being spectators. Your mentor works with a small group, so nobody gets left behind quietly.
Covered
Creative AI: art, music, storytelling
AI is an instrument, and instruments reward taste. Anyone can generate a thousand images; the skill is knowing which one is good and why. You will make work you would actually put your name on, and argue about where the line sits between using a tool and outsourcing your voice.
Covered
Responsible AI and life online
This is not a lecture about being careful. It is the practical stuff: what an app does with the photo you upload, how a deepfake is made and how it is spotted, why a model trained mostly on English gets your country wrong. Discussed with your group, not preached at you.
Covered
Demo day
You finish your own project and present it to a room: what you built, what broke, what you would do next. Standing up and explaining your choices is the part that stays with you long after the tools have changed — and it is the reason parents tend to notice a difference.
Covered
A personal AI project showcase: a working project you chose, built with your mentor and finished, presented at demo day to an audience — with an honest account of what broke, what you fixed, and what you would build next.
Offered in
Questions
My child already spends too much time on screens. Is this not more of the same?
It is the opposite of the same, and the distinction is the whole point of the track. Scrolling and building are not two versions of one activity: one is designed to hold attention, the other demands it. Your child already uses AI daily — that is not in question — and the only real choice is whether they understand what it does. A teenager who has built something and had to explain it to a room relates to their phone differently afterwards. The sessions are mentor-led with small groups and time away from the keyboard, and demo day is the visible proof: they will show you something they made, not something they watched.
Will this teach them to cheat on their homework?
They already know how — that skill required no course. What they generally do not know is that the model invents sources, that a teacher can usually tell, and that work handed in without being understood leaves a real gap at exam time. We address this directly rather than pretending it away: there is a session on using AI for schoolwork honestly, and it is a discussion, not a lecture, because moralising to a sixteen-year-old about a tool they use daily achieves nothing. In our experience the students who learn how the model actually works become markedly less impressed by its output, and more careful with it.
They have never coded, and I do not want them discouraged. Is six weeks realistic?
Yes, because the goal is not to produce a software engineer in six weeks — that would be a promise worth distrusting. The goal is one finished project they built and can explain. There is no prior coding requirement, mentors work with small groups precisely so that nobody drops out quietly, and the projects start small enough to finish. What discourages teenagers is not difficulty; it is being alone with something broken and nobody noticing. That is the failure mode we design against. If they finish and want more, the other modules are there — and if they finish and decide this is not their thing, they still leave able to explain how the technology works, which at their age is not a small return.
Other modules
Ready to take this module?