Ages 6–14 · Three levels · New

Your kid already uses AI. We make them smarter than it.

The AI Creator Program — a complete ladder for ages 6–14. From "what is AI?" to a live bot on the internet: three year-long courses where kids train real models, build real products, and learn the one skill schools don't teach — how to stay in charge of the machine.

✓ Ages 6–14 ✓ Real AI tools ✓ No prior experience required

Start with a trial lesson — your child builds something small and gets an honest level recommendation

Book trial lesson →

Why learn AI

Why your child should learn AI now

The tech world split kids into two groups: those who consume AI, and those who command it. Here's how we move yours to the second group.

📱

They already use it — unarmed

Your child meets AI daily. The question isn't whether; it's whether anyone taught them the rules. We teach safety from lesson one.

✍️

Unsupervised AI makes kids dependent

Asking ChatGPT for every answer stops the thinking. Our covenant makes AI explain and quiz — never write for them.

💼

The divide is forming now

Jobs won't disappear — they'll go to people who command AI instead of competing with it. That gap opens in childhood.

🔨

Consumption isn't the same as skill

Every lesson ends with something your child MADE: a trained model, a game, a book, a bot. 32 artifacts a year.

💡 Screens aren't the enemy — consumption is. We turn the same hours into building.

What they learn

What students will learn

Real tools at every age — with training wheels on the infrastructure, never on the thinking.

What AI actually is, how it learns from data, and why it makes mistakes
How to train real machine-learning models — images, sound, and gestures
Prompt engineering done properly, not guesswork
How to build AI into their own games and apps
Python, professional AI APIs, and deploying a bot to a real server (ages 12–14)
How to measure a product honestly — metrics, bias audits, and "what it can't do yet"
The safety rules: check everything, real people first, tell a grown-up

💡 Google's Teachable Machine at six. Machine Learning for Kids and Scratch at nine. Python and professional AI APIs at twelve. No toy simulators — the real thing.

Course structure

One ladder. Three rungs. Zero prerequisites.

Each level is a full year: 4 modules × 8 lessons = 32 lessons, ending in a family show.

Ages 6–8 · AI Explorers

They won't just play with AI. They'll train it.

What AI is, how it learns, and how it gets things wrong
An AI-illustrated storybook they author themselves
First games in Scratch with "smart" elements
A graduation invention, presented at the Invention Fair

They leave with: a published book, trained models, their first games, an inventor's diploma.

Ages 9–11 · AI Investigators

Real models. Real clients. Real numbers. Age nine.

Machine-learning models inside their own games
A chatbot built for a real client, delivered on Client Day
Data, confidence, and bias — measured, not argued
A capstone with a bias audit and honest metrics, pitched at Demo Day

They leave with: an AI-powered game, a delivered client bot, an audited product, a portfolio.

Ages 12–14 · AI Creator

Real Python. Real APIs. A bot that's live while they sleep.

Python from zero to confident, with dev-team culture
Code that commands professional AI models — personas, memory, guardrails
A Telegram bot deployed to a real server with real users
A capstone product with documentation, metrics, and a recorded demo

They leave with: a live bot, a shipped product, a demo video, a portfolio that opens doors.

💡 Entry at any age starts at that level's Module 1. New groups launch every 2–4 weeks — your child's year runs from their start date, not a school calendar.

Projects

What your child actually builds

Artifacts, not certificates. A certificate says a child attended. A live bot says a child built.

📖

An AI-illustrated storybook

Ages 6–8 author and publish a real book, illustrated with AI they direct themselves.

🎮

Games with real ML inside

Trained models become game mechanics — the camera, the voice, and the gesture are the controller.

🤝

A chatbot for a real client

Ages 9–11 take a brief, build a bot, and deliver it to an actual person on Client Day.

🚀

A live Telegram bot

Ages 12–14 deploy to a real server. It answers real users while they sleep.

🔍

A bias audit

Before anything launches, students measure where their model fails — and write it down honestly.

🎬

A capstone with a demo video

The year ends with a shipped product, its metrics, and a recorded demo for the portfolio.

Why Geeklama

Why Geeklama kids end up smarter than the machine

1 · The covenant

AI explains, quizzes and reviews our students' work — it never does it for them. Enforced warmly, every lesson, with the Explain Test: any line you can't explain, you delete or learn.

Parents tell us it changes how their kids use AI at school.

🛠️

2 · Real tools, right-sized

Teachable Machine at six. Machine Learning for Kids and Scratch at nine. Python, AI APIs and Telegram at twelve. No toy simulators.

🔒

3 · Safety you can inspect

No student accounts at 6–8. School-managed logins at 9–11. Server-side API keys and a privacy launch checklist at 12–14. Ask us anything.

🎭

4 · A show every 8 weeks

You don't get progress reports. You sit in the audience while your child presents — and handles a live glitch like a professional.

👥

5 · Small groups, hands-on

Groups cap at eight and every child builds every lesson. 1-on-1 available.

📊

6 · Honesty as a feature

Students measure with real metrics, write "what it can't do yet" sections, and run bias audits before launching.

🪜

7 · A ladder, not a course

96 lessons across three years, with bridges between every level — and a junior teaching assistant track for graduates at 14+.

💡 We're not raising kids who trust AI — we're raising kids AI has to answer to.

Format

How lessons work

💻

Live online classes

Interactive sessions with a live teacher — not pre-recorded videos.

👥

Mini-groups or 1-on-1

Groups cap at 8 students. Every lesson is hands-on for every child.

⏱️

45–55 minute sessions

45 minutes for ages 6–8, 55 minutes from age 9. Long enough to build, short enough to stay sharp.

💡 Everything runs in Chrome — nothing to install. A webcam is needed; from age 12, a real keyboard matters.

Pricing

Simple, transparent pricing

The AI Creator Program uses the same pricing as every Geeklama course — mini-groups or 1-on-1, billed by lesson package.

See pricing & packages →

Questions

Frequently asked questions

Is my child too young for AI?+

Too young to train a model with a teacher's help? Our six-year-olds do it in week three — no accounts, nothing stored, and safety rules they'll happily teach you. Too young to meet AI unsupervised — that's the actual risk, and it's already on your tablet.

Will this teach my kid to cheat with ChatGPT?+

The opposite. Our students learn to use AI as a tutor and to be honest about it. The Explain Test makes ghostwriting pointless: you present only what you can explain. Any line a student can't explain, they delete or learn.

My child has zero coding experience. Is that a problem?+

Not at all. The ladder has three entry points and no prerequisites at any level. A trial lesson tells you honestly which rung fits your child. Children who already code get bonus tracks rather than boredom.

What equipment do we need?+

A computer with Chrome and a webcam. Everything runs in the browser — nothing to install. For ages 12–14 a real keyboard matters; tablets and coding don't mix.

What if we miss lessons or start mid-year?+

New groups start every 2–4 weeks and your child's year runs from their own start date, not a school calendar. Catch-up lessons exist. The programme bends; the quality doesn't.

What exactly does my child have at the end?+

Artifacts, not certificates: a published storybook and trained models (6–8), a delivered client bot and an audited product (9–11), a live Telegram bot and a capstone with a demo video (12–14). There is a diploma too — but the portfolio is the point.

AI tools change every month — won't this be outdated?+

Tools change; the thinking doesn't. Data → model → verification, honest metrics, guardrails, ethics — that's what we teach. Our graduates will command whatever ships next year, because they understand what's inside it.

What's the workload for parents?+

About ten minutes a week: hear the pitch, play the game, answer as a beta tester. Our parent messages tell you exactly what to ask. The building is 100% the child's — that's the covenant.