Your privacy choices
We use cookies that are needed to run the site, and — only if you agree — cookies that measure how the site is used and how our ads perform. You can change or withdraw your choice at any time from “Cookie preferences” in the footer. Our Privacy Policy lists every tool we use and how long we keep the data.
Not a chatbot demo. Five steps, from using AI to directing it.
Your child learns AI the way they learn to code: by building it. They train an image model by Level 3, build a miniature text generator in Level 7, run a private AI on the class laptop in Level 8, and finish with an AI project of their own. Teens carry on in an AI and data path.
Our AI classes for kids ages 7 to 14 are part of Junior Creator, an in-person coding program. Students train an image model in Level 3, build a miniature text generator in Level 7, run a private AI on the class laptop in Level 8, and finish with an AI project of their own in Level 13. Teens ages 14 to 18 carry on in Ultimate Innovator's AI and data path.
AI runs through our coding programs rather than being sold on its own, because a child who can code is the one who can tell when an AI has got it wrong.
7 to 14
Five steps across thirteen levels: using AI, writing the rules a game character follows, training models on data they collected, building a text generator, and directing an AI assistant.
Junior Creator, coding and AI for ages 7 to 14 →14 to 18
The AI and data path: what a model is and how one gets trained, then datasets, fine-tuning and generative AI, then machine learning, finishing able to say what their model does badly and why.
Ultimate Innovator, engineering for ages 14 to 18 →Your child always knows which step they are on, and can tell you the difference between the five.
Their project uses AI someone else built: speech and translation, from Level 1.
They write every rule the brain follows: enemies that sense, decide and act, and their own chatbot.
No rules written. The machine learns from examples your child gathered, then shows what the data left out.
They build the thing that predicts and creates: a miniature text generator, which is how the big ones stop being magic.
They direct an AI assistant, then verify, test and take responsibility for its work before trusting it.
The Junior Creator levels where your child builds or trains an AI themselves. The levels between them build the coding those projects stand on.
Functions, lists and generative art, plus a real image model they train and then play at rock, paper, scissors.
Physics, particles and procedural worlds, plus a trainable neuron that learns in front of them.
Files, data and charts, and a miniature text generator they build themselves from counted patterns.
Machine learning end to end: classifiers, a neuron, hand tracking, and an AI running on the class laptop.
Real interactivity: events, design systems, saved data, and a page controlled by a model they trained.
Web apps with state and live data, plus neural networks running inside the browser they built.
One signature AI product of their own invention: specify, build, test, evaluate and pitch it.
Real, and deliberately honest. Students train models on data they collect, write a classifier and a neuron themselves, build a miniature text generator, and in Level 8 run a large language model on the classroom laptop and check it for bias.
The AI tools run on the classroom laptop rather than the cloud, so no child's camera image, voice or writing leaves the room. Every tool is free, and families buy nothing.
We never call an if-then block real AI. We teach your child the difference, so that when they say they built an AI, they can tell you exactly what they mean.
19 centres across Ontario, British Columbia and Alberta. Schedules and prices are set by each centre and shown on its own page.
Seven, in Junior Creator. Their first projects use AI someone else built, like speech and translation, and by Level 3 they are training an image model of their own. At five to seven, Junior Thinker builds the groundwork: clear instructions, noticing what went wrong and trying again. A five-year-old isn't building AI, and we won't tell you otherwise.
Real, and deliberately honest. Students train models on data they collect, write a classifier and a neuron themselves, build a miniature text generator, and in Level 8 run a large language model on the classroom laptop and check it for bias. We never call an if-then rule real AI, and we teach the difference explicitly.
Because AI produces code but can't judge it. Deciding what to build, spotting what is subtly wrong and knowing what good looks like all come from having built things. Children who code end up directing AI; children who don't are left trusting it.
No. The AI tools in Junior Creator run on the classroom laptop rather than the cloud, so no child's camera image, voice or writing leaves the room.
No. Junior Creator starts from nothing, and the AI work comes in the same order as the coding: a child writes rules before they train a model, and trains a model before they build one.
Yes. Ultimate Innovator, for ages 14 to 18, has an AI and data path. Artificial Intelligence starts from what a model is and how one gets trained; Machine Learning goes deeper and is usually taken after Python Advanced.
Junior Creator and Ultimate Innovator both run online, as live classes with the same instructor-to-student ratio as a class in a centre.
Sixty minutes in a real class. Your child builds something, and the instructor tells you honestly which level fits.