AI Builders.
The top of the ladder. Teens stop being users and become makers: they run as start-ups, write system prompts robust enough to survive attack, produce real media, argue AI ethics with evidence, and ship a complete project from a blank page to a public exhibition.
Six things a Builder can actually do.
Not topics covered — capabilities they leave holding.
Run a research pipeline
Run a full research-to-artifact pipeline with real analysis — Gemini Deep Research and NotebookLM — from a framed question to a defensible position.
Write robust system prompts
Write system prompts sturdy enough to survive attack, and deploy them on a physical AI agent that a rival team will try to break.
Produce real media
Produce media through professional AI pipelines — script it, generate it, cut it, score it — and ship a finished piece.
Argue AI ethics
Argue both sides of a real AI-ethics question with evidence — not opinions, sourced positions they can defend under pressure.
Ship a complete project
Ship a complete project, A to Z, from a blank page to something presented publicly to a room of people.
Compare universities (Grade 12 track)
Hold an evidence-based university and major comparison — answered with sources, never by an AI verdict.
From smart prompting to teaching the teachers.
Fourteen sessions in five moves. The class starts as individuals and, halfway through, becomes a set of AI companies with rotating roles.
One skill per session.
Six sessions, six skills: Smart Prompting, The Investigation Room, The Study Lamp, The Writer's Desk, The Strategy Table, Maker Mindset. Outcome: prompts with full RTCF structure across study, writing and strategy — one artifact per skill.
- Session 1
Smart Prompting
The RTCF structure — role, task, context, format — that every later skill is built on.
- Session 2
The Investigation Room
Ask a machine and interrogate what comes back instead of accepting it.
- Session 3
The Study Lamp
Prompting turned on their own studying — RTCF applied to learning.
- Session 4
The Writer's Desk
Directing a machine through a real piece of writing, keeping their own voice.
- Session 5
The Strategy Table
Prompting for planning and decisions, not just answers.
- Session 6
Maker Mindset
The turn from consuming to producing — the bridge into the rest of the course.
Real research, real analysis.
Frame a real question, gather, interrogate sources, compare, hunt contradictions, and form a defensible position. Grade 12s run the same method on universities and majors.
Frame & gather
Turn a vague curiosity into a real question, then gather real sources with Gemini Deep Research and NotebookLM.
Interrogate & take a position
Compare sources, hunt contradictions, and land on a position they can defend — with citations, not vibes.
The AI Company opens.
Machine-learning day. The class stops being individuals and starts operating as companies with rotating roles — CEO, prompt engineer, fact-checker, ethics officer, designer, pitcher.
How machines learn
A 70-year history timeline, the AI / ML / deep-learning / chatbots nesting cups, and the three ways of learning — laid out so it finally clicks.
Companies & a live agent
Roles are assigned, the companies form, and the session closes on a live agent demo — the thing they'll soon build and defend.
The Studio.
Master and system prompting, visual identity, then production. Each company writes a persona and loads it into the physical Zakoot robot — and a rival team tries to break it.
Session 10 · Persona & the robot
Master and system prompting plus a visual identity. Each company writes a persona, loads it into the physical Zakoot robot, and defends it while a rival team tries to break it.
Session 11 · The 60-second video
Produce a real 60-second instruction video — script in Gemini, generate in Veo, cut in CapCut, score in Suno — with a warm-Lebanese-Arabic quality check. The best one actually ships to the Explorers level.
AI on Trial.
The ethics block, staged as a full courtroom.
Three charges
It lied. It copied an artist. It took a job. Three real accusations, each argued on its merits.
Evidence packs
Prosecution and defence each get an evidence pack. Teams argue both sides — no easy verdicts.
The closing question
It closes on cognitive debt: what am I outsourcing — and what does it cost me to hand it over?
From a blank page to a shipped artifact — then teaching adults.
From a framed question to a shipped artifact: a report, a dashboard, a quiz app, a chatbot agent, or a 2D game. Then the strongest proof the programme works — students teaching adults.
Session 13 · A to Z
Each company takes a framed question all the way to a finished artifact — report, dashboard, quiz app, chatbot agent, or 2D game.
Session 14 · Kids Teach the Teachers
In the finale, each company runs a 15-minute AI mini-workshop for invited teachers — the strongest proof the programme works: students teaching adults.
The most advanced level.
Builders
14 sessions · small group · runs as start-ups
Enquire about a placeGrade 12s get the university & major research track — an evidence-based comparison answered with sources, never an AI verdict.
Watch them ship it.
Join the next Builders cohort, or bring the level into your school.