02 //Academy
- Data
- Models
- Testing
- Judgement
How machines learn.
AI is explored by using it and taking it apart: training something simple, seeing where it gets things wrong, and asking why.
01 //What learners build
Four things, in order.
- 01How a model learns
Understand that what AI learns depends on what it is shown.
Showing a computer examples, then seeing what it does with something new. Learners meet data, training, patterns and predictions — and why more examples don't always mean better results.
- 02Training something simple
Move from “AI can recognise things” to “I understand why.”
Learners collect examples, label them, train a simple model and test it on examples it hasn't seen. They discover that a model which works on familiar examples may still struggle with something new.
- 03Where it goes wrong
Question an AI result instead of believing it automatically.
What happens when examples are incomplete, unbalanced, unclear or unlike the training data. Mistakes are made part of the lesson rather than hidden.
- 04Using AI thoughtfully
Know when to use AI, and when not to trust it unchecked.
Learners compare AI-generated answers with evidence, identify questionable outputs, and learn when a person needs to step in.
02 //The AI journey
Observe, train, test, question, use.
How a learner moves through the path
01
Observe
What does the machine notice? Learners explore patterns and predictions.
02
Train
What happens when we give it examples? Learners build a simple model from their own data.
03
Test
What happens when we show it something new? Predictions are tested against unseen examples.
04
Question
Why did it get this wrong? Learners investigate mistakes, data quality and bias.
05
Use
How can AI help without thinking for us? Learners apply it thoughtfully and verify its outputs.
03 //AI is not magic
It's a system we can investigate.
A learner sees an AI recognise an image and asks how it knows. That's where the learning begins.
Unpacking the process
01
Data
What did the system learn from?
02
Patterns
What relationships did it find?
03
Model
What did it build from those patterns?
04
Prediction
What does it think about something new?
05
Evaluation
Was it right? And then: why?
04 //What learners explore
The ideas underneath the tools.
- 01
Data — AI depends on examples and information
- 02
Models — how systems use learned patterns to predict
- 03
Training — how examples influence what a model learns
- 04
Testing — how a model behaves with information it hasn't seen
- 05
Accuracy — a system can be right sometimes and wrong sometimes
- 06
Bias — how training examples influence results
- 07
Generative AI — systems that produce text, images and ideas
- 08
Verification — why AI-generated information should be checked
- 09
Responsible AI — privacy, appropriate use and human judgement
05 //Learning through experiments
Every concept should have something to investigate.
- 01Experiment 01
Teach the machine
Give a simple model examples and see whether it recognises a new one. Data, categories, training and prediction.
- 02Experiment 02
Can it see the difference?
Change the examples and observe how the predictions change. Training data and model behaviour.
- 03Experiment 03
Break the model
Give it examples that don't represent the real world well, and watch what happens. Data quality, bias and limitations.
- 04Experiment 04
Human against AI
Give a learner and an AI the same problem. Compare the answers. Which is better, and why?
- 05Experiment 05
Ask better questions
Try different ways of communicating with a generative system and compare the outputs. Instructions, context, specificity.
- 06Final challenge
Investigate an AI system
Choose a tool or experiment and investigate it: what it does, what it learns from, where it works, where it struggles, and how a person should use it responsibly.
06 //The learner progression
From “AI is magic” to “I can investigate it.”
- 01Explorer
“What can AI do?”
Discovering AI through simple, interactive experiments.
- 02Experimenter
“What happens if I change the examples?”
Exploring training data and model behaviour.
- 03Investigator
“Why did it get this wrong?”
Studying errors, bias and limitations.
- 04Critical user
“Can I trust this answer?”
Learning to verify AI-generated information.
- 05AI thinker
“How should we use this?”
Understanding that powerful technology also requires judgement.
07 //What parents can expect
AI becomes a thinking skill.
- 01
Curiosity
Learners become comfortable asking how technology works.
- 02
Critical thinking
They learn not to accept computer-generated answers automatically.
- 03
Data awareness
They begin understanding that the quality and source of information matter.
- 04
Problem solving
They experiment, test ideas and improve their approaches.
- 05
Digital responsibility
Early awareness of privacy, bias and responsible technology use.
- 06
Confidence
Instead of being intimidated by AI, learners explore it thoughtfully.
08 //What learners leave with
Don't teach children to use AI. Teach them to understand it.
The tools will change, the models will change, the interfaces will change. The ability to question, experiment, evaluate and think critically will remain valuable — which is why this journey begins with curiosity rather than hype or fear.
- An understanding of data and models
- And experience training and testing something simple.
- An awareness of bias and limitations
- And the habit of checking what an AI produces.
- A foundation for responsible use
- “I don't just use AI. I know how to question it.”
Other paths
Other paths
- 01
Coding
Coding starts with logic a learner can see and moves into written code, so the step from blocks to a real language feels like the next sentence rather than a new subject.
- 02
Robotics
Robotics connects a program to the physical world, where a mistake is visible immediately: the motor turns the wrong way, the sensor reads nothing, or the robot simply refuses to do what you told it to.
- 03
Game Development
Game development turns a familiar pastime into a design problem: rules, movement, difficulty and feedback all have to be decided by the person making it.
Use it, question it, learn from it
The most important capability is still human.
Knowing what question to ask, what evidence to look for, and when a person needs to decide.
