The Complete AI & Machine Learning Curriculum for Kids (Ages 5–10)

A free, parent-led journey through artificial intelligence — from sorting socks to training neural networks. 60 phases, 360 lessons and 60 hands-on projects across three age-banded courses. Here is everything your child will learn.

← Back to home Play the AI games

Ages 5–6

Little Explorers — AI for Kids Ages 5–6

120 play-first lessons across 20 phases — from 'what is smart?' all the way to training real AI models, almost entirely screen-free.

At this age children learn through play, stories and repetition, so every lesson is a 10–15 minute game you lead, with the exact words to say. Sessions work best 2–3 times a week; one phase takes about two to three weeks. Don't chase speed — chase giggles and 'again!'. Re-playing an old lesson counts as progress too. The course is long on purpose: it's one to two years of gentle companionship, growing with your child.

Start the Little Explorers course →

Phase 1: What does 'smart' mean?

Your child discovers that smart isn't knowing everything — smart is being able to learn. And some special machines can learn too.

People learn, toys don't
Learning needs practice
Meet a learning machine
Robot Mummy / Robot Daddy
Brains and computer brains
The learning-machine hunt
Project: Project: Train-a-Grown-Up Week

Phase 2: Sorting & groups

Classification — putting things into the right group — is the first job of real AI systems. Your child becomes a sorting expert with toys, snacks and socks.

The big sort
Two-box machines
Tricky ones
Sort with your eyes closed
Sorting by two things at once
The sorting relay
Project: Project: The Sock Sorting Machine

Phase 3: Patterns everywhere

Pattern-spotting is the muscle every learning machine flexes. Your child trains it on shapes, sounds, days and stories until patterns jump out of the world by themselves.

What comes next?
Pattern detective at home
Sound and clap patterns
Fortune teller: patterns in our day
When predictions go wrong
Story patterns
Project: Project: My Pattern Book

Phase 4: Predictions & good guesses

From spotting patterns to using them: your child learns to make predictions, attach confidence to them, and update guesses when new clues arrive — the thinking style of every forecasting machine.

Guess the gift
How sure are you?
The weather guesser
Predict the people you love
Changing your mind is smart
The prediction carnival
Project: Project: The Family Prediction Tournament

Phase 5: Robot eyes: how machines see

Cameras grab dots; learning machines work out what the dots mean. Your child takes apart 'machine seeing' with zoomed photos, pixel art and peek-a-boo experiments.

Eyes vs cameras
Make a picture from squares
What makes a cat a cat?
Peek-a-boo for robots
Tricky light
Robot eye test
Project: Project: The Pixel Portrait Gallery

Phase 6: Robot ears & robot touch

Microphones turn sound into guesses; sensors turn the world into numbers. Your child explores machine hearing and touch — and the senses machines simply don't have.

Ears vs microphones
Sound detective
The talking game (speech to action)
Robot touch and other sensors
Senses machines don't have
Build a five-senses robot map
Project: Project: Build 'Beep', the cardboard robot

Phase 7: Giving instructions (little algorithms)

Before machines learned, they obeyed. Your child discovers what instructions really are: exact, ordered steps — and that computers follow them with zero imagination.

The jam sandwich robot
Order matters
The treasure-hunt program
Loops: do it again and again
If this, then that
When rules run out
Project: Project: The Family Instruction Book

Phase 8: Learning from examples (training!)

The heart of the course: machines learn the un-writable by seeing many examples. Your child runs trainings, discovers why more and varied examples win, and meets the train-then-test rhythm.

The example game
More examples, better learning
Bad examples make bad learners
All kinds of examples
Train, then test
You are the trainer
Project: Project: The Training Academy

Phase 9: Learning from mistakes (rewards & tries)

A second way machines learn: not from labelled examples, but from trying, failing, and chasing rewards — the warm-cold game that trains game-playing AI and robots.

Warmer, colder
The reward game
Try, fail, try different
Practice makes patterns
Two ways to learn
Coach a clumsy robot
Project: Project: The Backyard Robot Olympics

Phase 10: Data all around us

Examples, feedback, sensor numbers — it's all data. Your child learns to collect it, picture it, and read its stories: the first full lap of the data-science loop, in crayon.

What is data?
The great count
Picture the numbers
Ask the data a question
Machines eat data
My data day
Project: Project: The Family Data Museum

Phase 11: Labels & names

A label is the answer glued to an example — and labelling is teaching. Your child becomes a careful labeller and learns why sloppy names make confused machines.

Naming is teaching
The label factory
Wrong labels, wrong world
Hard to label
The label rule book
The labelling party
Project: Project: The House Museum of Labels

Phase 12: Memory machines

Machines remember differently from people: exactly, hugely, and only what's saved. Your child explores storing, finding and forgetting — and who decides what gets remembered.

Where machines keep things
Finding fast
When the box is full
Exact memory, fuzzy memory
Does the machine remember me?
Forgetting on purpose
Project: Project: The Family Memory Library

Phase 13: Talking machines

Machines that finish sentences can chat, rhyme and answer — without understanding or feeling any of it. Your child learns to enjoy talking machines and see straight through them.

The sentence-finishing machine
The clever parrot
Ask it anything?
When the machine makes things up
Kind to machines?
Our talking-machine rules
Project: Project: Polly's Question Show

Phase 14: Machines that make things

Some machines don't just sort and answer — they MAKE: pictures, rhymes, stories. Your child discovers where machine creativity comes from (remixed patterns) and where human creativity stays ahead (the idea, the choice, the why).

Where do new pictures come from?
Be the making machine
The style copier
Whose picture is it?
The idea is yours
Made-up music and rhymes
Project: Project: The Remix Art Show

Phase 15: Robots that move

A robot is a machine that senses, decides and ACTS in the world. Your child runs the sense-think-act loop in their own body — and discovers why the easiest human things are the hardest robot things.

Sense, think, act
Bump and turn
Why robots fall over
Robot hands
Robots at work
Design your robot
Project: Project: The Human Robot Show

Phase 16: When machines get it wrong

Machines fail — confidently, strangely, fixably. Your child learns to laugh at machine mistakes, diagnose their causes, and prescribe the cure: better teaching.

Everyone makes mistakes — machines too
The funny fails collection
Why did it fail?
The 'not sure' superpower
Whose job is checking?
Fixing by teaching
Project: Project: The Machine Hospital

Phase 17: Fair and kind machines

Machines should work for EVERYONE — and data about people deserves asking first. Your child meets fairness, bias and consent at the height of a five-year-old's fierce sense of justice.

That's not fair!
The cookie machine
The everyone test
Asking first
Sharing kindly
The kind machine rules
Project: Project: The Fairness Fix-It Crew

Phase 18: Real or pretend?

Machines can now make pictures and voices that look and sound real. Your child builds the gentlest, sturdiest defence: a calm checking habit — not fear, not gullibility, but the question.

Pictures can pretend
Voices can pretend too
Stories, jokes and tricks
Check with a grown-up
The careful believer
Real-or-pretend in the wild
Project: Project: The Real-or-Pretend Game Show

Phase 19: AI helpers in our world

A grand tour of where learning machines actually help — homes, hospitals, farms, roads — and the people behind every one of them. The course's worldview lesson: helpers, built and checked by humans, working FOR us.

Helpers hiding at home
Helpers that keep us healthy
Helpers on farms and in shops
Helpers on the road
The people behind the machines
Helper or boss?
Project: Project: The Neighbourhood AI Safari

Phase 20: My first real AI — graduation

Everything converges: your child trains, tests, breaks and fixes a REAL machine-learning model with Teachable Machine — then graduates as a Little Explorer who has personally done what most adults never will.

The grand revision quiz
Meet the real training machine
Training day
Test it fairly, break it joyfully
Fix it by teaching
Rehearsal for the show
Project: Graduation: The Toy Recogniser Show & Diploma Ceremony

Ages 7–8

Junior Builders — AI for Kids Ages 7–8

120 lessons across 20 phases — real vocabulary (data, features, training, bias), a lab notebook, hands-on experiments, and AI your child trains, tests and improves alone.

Seven- and eight-year-olds can read simple text, follow multi-step instructions, and love being treated as scientists. Lessons run 15–25 minutes; read them together at first, then let your child lead while you assist, fetch props and ask 'why do you think that happened?'. Start a cheap notebook as the AI Lab Notebook — writing things down roughly doubles retention and becomes a keepsake. Computer lessons use free tools (Teachable Machine, Google Sheets) with no child account; you supervise and control any uploads. The course is long on purpose: a year-plus of real apprenticeship that grows with your child.

Start the Junior Builders course →

Phase 1: Rules vs learning

The single most important distinction in computing: ordinary programs follow rules people wrote, while machine-learning programs write their own rules from examples. Your child learns to tell them apart on sight.

Be a computer for ten minutes
Some problems have no rules
Spot the learner
The instruction language
Why learning beats rules (sometimes)
The two-machine demonstration
Project: Project: The Rule-Bot vs Learn-Bot Challenge

Phase 2: Be the algorithm

An algorithm is a precise recipe that always finishes. Your child writes, runs and debugs real algorithms by hand — and discovers that a clever algorithm beats raw effort.

What makes a recipe an algorithm
The sorting race
The guessing-game algorithm
Debugging detective
Algorithms with choices
Human computer relay
Project: Project: The Algorithm Olympics

Phase 3: What is data?

Data is recorded facts — the food every learning machine eats. Your child learns what counts as data, collects it deliberately, and starts the lab notebook habit that defines a data scientist.

Data is everywhere
The survey
Tables: rows and columns
Good data, bad data
Counting and comparing
The lab notebook
Project: Project: The Family Data Study

Phase 4: Collecting & labelling datasets

Supervised learning needs labelled examples — data with the answers attached. Your child builds real labelled datasets and discovers why careful, consistent labels make or break a machine.

Labels: the answers attached
Build a labelled dataset
When labels disagree
How much data is enough?
Cleaning the data
The dataset card
Project: Project: The Custom Dataset Build

Phase 5: Features: what the machine looks at

A feature is a measurable property a machine uses to decide. Choosing good features is half of machine learning — and your child learns to find the features that separate one thing from another.

What is a feature?
Good features separate things
Features in numbers
Too many, too few
Hidden features
Feature detective challenge
Project: Project: The Great Sorting Machine Design

Phase 6: Training & testing

Your child trains a real machine-learning model in the browser — and learns the golden rule that protects all of machine learning: never test on what you trained on.

Meet Teachable Machine
Train your first model
The golden rule of testing
Reading the scores
Improving the model
Train an ears model
Project: Project: The Pet Detector (or Toy Detector)

Phase 7: Computers that see

Vision models turn pictures into predictions. Your child opens up machine seeing — pixels, features, layers — and discovers exactly where and why it breaks.

Pictures are numbers
From dots to features
Why vision breaks
Fooling the model
Vision in the world
Design a vision system
Project: Project: The Vision Lab Report

Phase 8: Computers that hear & speak

Audio models turn sound into meaning and back. Your child explores the speech pipeline — sound to words to action — and discovers why talking to machines is harder than it looks.

Sound is a wave of numbers
Recognising sounds
From sound to words
The speech pipeline
Why machines mishear
Design a voice helper
Project: Project: The Family Voice Assistant (acted & designed)

Phase 9: Kinds of learning

Machines learn in three great ways: from labelled examples, from unlabelled data, and from rewards. Your child maps the three families and learns to recognise which is which.

Learning with answers (supervised)
Learning without answers (unsupervised)
Learning from rewards (reinforcement)
Which kind is it?
The right tool for the job
Build one of each
Project: Project: The Three Learners Exhibition

Phase 10: Prediction machines

Many AI systems are really prediction machines: they spot patterns in past data to guess what's next. Your child learns how machines forecast, and where forecasts go right and wrong.

Predicting from patterns
The trend line
More clues, better predictions
When predictions fail
Predicting people
Build a predictor
Project: Project: The Family Forecaster

Phase 11: Clustering: finding hidden groups

Sometimes data has no labels and the machine must discover the groups itself. Your child learns unsupervised clustering by hand — finding the natural categories no one named.

Groups nobody named
Many ways to group
How close is close?
How many groups?
Spotting the odd one out
Cluster a real dataset
Project: Project: The Discovery Study

Phase 12: Decision trees you can draw

A decision tree is a real machine-learning model simple enough to build by hand: a flowchart of yes/no questions ending in answers. Your child builds, tests and improves their own.

The question tree
Good questions split well
Building it from data
Testing the tree
Trees vs other models
The expert system
Project: Project: The Recommendation Tree

Phase 13: When AI gets it wrong

Every model fails. Your child becomes an error analyst — diagnosing why models go wrong, distinguishing data problems from model problems, and prescribing the right fix.

Rubbish in, rubbish out
Data problem or model problem?
The confusion detective
Confidently wrong
The fix-it loop
The error report
Project: Project: The Model Hospital

Phase 14: Fairness & bias

AI that learns from biased data treats people unfairly — and it can hurt real people. Your child learns to spot bias, understand where it comes from, and run a real fairness audit.

When the data leaves people out
Where bias comes from
The everyone test
Fixing unfairness
Who decides what's fair?
The fairness charter
Project: Project: The Fairness Audit

Phase 15: Chatbots & language machines

Language models predict the next word — and that single trick, at huge scale, produces machines that seem to talk. Your child builds chatbots by hand and learns how the big ones really work.

The next-word machine
Build a paper chatbot
Context changes everything
Fluent isn't the same as true
What it knows and doesn't
Using language machines wisely
Project: Project: The Chatbot Investigation

Phase 16: Generative AI

Some AI doesn't sort or predict labels — it MAKES new things: pictures, text, music. Your child learns how generation works (remixing learned patterns) and the questions it raises.

Machines that make
Be the generator
Style and the how
Whose work is inside?
Real or generated?
Generative AI for good
Project: Project: The Generation Exhibition

Phase 17: AI safety & trust

How much should we trust a machine, and how do we use AI safely? Your child builds a calibrated trust-meter, learns verification and privacy habits, and writes a personal AI safety code.

The trust-o-meter
Verify what matters
Keeping yourself private
When machines decide about people
Being kind in an AI world
My AI safety code
Project: Project: The Family AI Safety Guide

Phase 18: AI around town

A field guide to where AI actually lives in the world — and the people who build, train and check every system. Your child connects everything learned to real applications and real jobs.

AI in one ordinary day
AI that keeps us healthy
AI that moves the world
AI in science and discovery
The people who build AI
AI's big questions
Project: Project: The AI Field Guide

Phase 19: Plan like an engineer

Engineers don't start by building — they plan: define the problem, the data, the test, the risks. Your child learns the discipline that turns a vague idea into a buildable project, preparing for the capstone.

Start with the problem
Plan the data
Plan the test
What could go wrong?
The project proposal
Build small, test often
Project: Project: The Capstone Proposal

Phase 20: Build it — the AI Science Fair

Everything converges: your child executes their proposal — building, testing, and improving a real AI project — then presents it at a science fair, graduating as a Junior Builder.

Set up the build
Build and grow
Test it properly
Improve it
Prepare the presentation
Reflect on the journey
Project: Graduation: The AI Science Fair

Ages 9–10

Young Creators — AI for Kids Ages 9–10

120 lessons across 20 phases — under the hood at last: algorithms, neural networks, language and generative AI, ethics and an engineering capstone with a real Demo Day.

Nine- and ten-year-olds can read independently, reason abstractly, and use a computer with supervision — so here the child reads the lessons themselves and you become a discussion partner and safety net (you control all accounts and uploads; Teachable Machine, Scratch and Google Sheets need no child account). Sessions run 25–40 minutes. This course genuinely opens the box: real algorithms, the training loop, how neural networks and language models actually work, AI ethics, and a multi-session capstone. Keep the lab-notebook habit; by the capstone it becomes a real project journal. The depth is real — many university students would find this a solid foundation.

Start the Young Creators course →

Phase 1: Algorithms, precisely

An algorithm is a precise, finite procedure that solves a problem. Your child learns to define, analyse and compare algorithms — and discovers that cleverness, not effort, is what makes a procedure fast.

What an algorithm really is
Measuring efficiency
The power of halving
Two ways to get an algorithm
Inside the training loop
Algorithms all around
Project: Project: The Algorithm Olympics & Analysis

Phase 2: Programming meets machine learning

Your child writes real code in Scratch, then sees the deep contrast: code is rules a human wrote, while machine learning produces behaviour the computer derived from data. Two ways to make a computer act.

Hello, Scratch
Code is rules you wrote
ML is behaviour from data
When to code, when to learn
Build a real Scratch project
Code that uses a model
Project: Project: The Two-Ways Showcase

Phase 3: Data science I: spreadsheets

Spreadsheets are the data scientist's first real tool. Your child learns to structure, sort, filter and question data in a spreadsheet — turning rows and columns into answers.

The spreadsheet: rows and columns
Sorting and filtering
Counting and calculating
Features and the data map
Real datasets
The data question cycle
Project: Project: The Real Data Investigation

Phase 4: Data science II: charts & honesty

Charts make data speak — or lie. Your child learns to visualise data well, spot outliers and trends, and recognise the tricks that make honest-looking charts deceive.

Charts that show the truth
Outliers and what they mean
Finding the trend
How charts lie
Statistics that mislead
Tell an honest data story
Project: Project: The Data Journalism Report

Phase 5: Probability & uncertainty

AI lives in uncertainty — it deals in probabilities, not certainties. Your child learns to reason about chance, express confidence as numbers, and judge whether confidence is well-calibrated.

The language of chance
Combining chances
Updating with evidence
AI's confidence numbers
When confidence is wrong
Reasoning under uncertainty
Project: Project: The Uncertainty Investigation

Phase 6: Classification by hand

Your child builds two real classifiers — nearest-neighbour and decision trees — entirely by hand, understanding exactly how each decides. Real models, fully transparent, no magic.

Nearest neighbour: judge by company
Decision trees: twenty questions
Building a tree from data
Testing and overfitting
Comparing classifiers
What models can't do
Project: Project: The Classifier Workshop

Phase 7: The training loop

How does learning actually happen? Through a loop: predict, measure the error, adjust to reduce it, repeat. Your child runs this loop in detail and understands the engine inside every learning machine.

Learning as reducing error
Loss: measuring wrongness
Adjusting in the right direction
Learning rate: how big a step
When training goes wrong
The whole loop, end to end
Project: Project: The Learning Loop Simulator

Phase 8: Neural networks I

Neural networks are layers of simple deciders with adjustable connections. Your child builds the pieces — neurons, weights, layers — and acts out how a network computes, demystifying deep learning.

The artificial neuron
Weights are the knowledge
Layers: neurons working together
How a network computes
Training the network
Why neural networks are powerful
Project: Project: The Human Neural Network

Phase 9: Neural networks II: deep learning

Deep networks with many layers learn rich features — from edges to objects, from sounds to words. Your child explores what layers actually learn and why depth unlocked modern AI.

Why go deep?
What vision layers learn
The same idea, every sense
Data and compute: the fuel
The black box problem
Deep learning in the world
Project: Project: The Deep Learning Explainer

Phase 10: Language AI I: how it works

Language models predict the next word — and that, at vast scale, produces machines that write and converse. Your child learns how they really work, from words-as-numbers to next-word prediction.

Words as numbers
Next-word prediction
Learning from oceans of text
Context: reading the whole conversation
Does it understand?
Why fluent isn't true
Project: Project: The Language Model Explainer

Phase 11: Language AI II: using it wisely

Knowing how language models work, your child learns to use them well — prompting effectively, verifying outputs, protecting privacy, and keeping a clear head about a tool that talks.

Prompting: asking well
Verifying what it tells you
Privacy and language models
A tool, not a person
What it's good and bad for
My language AI guide
Project: Project: The Language AI User's Manual

Phase 12: Generative AI

Generative AI creates new images, text, music and more. Your child learns how generation works, explores deepfakes and discernment, and engages the genuine questions of creativity, credit and consent.

How machines generate
The creativity question
Deepfakes and discernment
Whose work trained it?
Using generative AI well
Generative AI showcase
Project: Project: The Generative AI Investigation

Phase 13: Reinforcement learning

Some AI learns by doing — trying actions, getting rewards, and improving through experience. Your child learns how reinforcement learning works, why self-play produces superhuman game AI, and where it fits.

Learning by reward
Designing the reward
Explore or exploit?
Self-play: getting superhuman
Why the real world is harder
Where reward-learning fits
Project: Project: The Reward Learner

Phase 14: Robotics

Robots join AI to the physical world through a sense-think-act loop. Your child learns why the physical world is AI's hardest arena, and what it takes for a machine to act in reality.

Sense, think, act
Why sensing is hard
Why acting is hard
Robots that learn
Robots among people
Design a robot
Project: Project: The Robot Design Lab

Phase 15: AI history

AI has a rich history of people, ideas, breakthroughs, and disappointments. Your child learns the story — from Turing's question to today's systems — and sees that AI was built by humans making choices.

Turing's question
Big dreams, hard reality
Two ways to build a mind
The breakthroughs
AI is made by people
Where might it go?
Project: Project: The AI Story

Phase 16: Ethics I: fairness & justice

AI makes decisions affecting real people at vast scale — so fairness becomes a matter of justice. Your child learns how bias scales, how to audit for it, and why fairness is a human-values question.

Bias at scale
Where injustice enters
Auditing for fairness
Fixing and preventing
Who decides what's fair?
A justice charter for AI
Project: Project: The Algorithmic Justice Report

Phase 17: Ethics II: safety & alignment

Powerful AI must be safe and aligned — doing what we actually want, not just what we literally said. Your child learns privacy, safety, the alignment problem, and writes a family AI code.

The alignment problem
Privacy in an AI world
Keeping powerful AI safe
Honesty, deception, and trust
Who's responsible?
The family AI code
Project: Project: The AI Safety & Ethics Guide

Phase 18: AI careers & how it's made

AI is built by many kinds of people doing many kinds of work. Your child learns the real roles, the process of making AI, and discovers that they've already practised much of the actual work.

The many jobs in AI
You've done the work
How AI gets made
What makes good AI work
Learning more, going further
Imagine your AI future
Project: Project: The AI Maker's Guide

Phase 19: Capstone planning

Before building, an engineer plans. Your child writes a real proposal for their capstone AI project — defining the problem, data, method, evaluation, ethics, and milestones — and defends it to a sponsor.

Choosing a worthy problem
Planning the data
Planning method and evaluation
Planning for ethics and safety
The proposal and milestones
Defending the proposal
Project: Project: The Capstone Proposal

Phase 20: Capstone build & Demo Day

Everything converges: your child builds, tests, and improves their capstone AI project from their proposal, then presents it at Demo Day — graduating as a Young Creator with a genuine foundation in AI.

Set up and build small
Build and grow
Test it properly
Improve it
Prepare for Demo Day
Demo Day & graduation
Project: Graduation: Demo Day