Course Beginner-friendly
How AI actually works
Not a mind, not a magic box, not a database of answers — an AI is a machine that learns patterns from examples and predicts. This course builds one from the idea up, so the headlines stop being magic.
Everyone talks about AI; almost no one can say what is happening inside one. This course takes the lid off — not how to use it, but the machine underneath. How a program that only does arithmetic can recognise a face, finish your sentence, or make a confident mistake. We build it from the idea up — learning from examples, the training loop, why it predicts rather than knows. By the end, an AI headline reads like a story you understand — not magic, and not doom.
What you'll be able to do
- Explain why AI learns from examples instead of following written rules.
- Trace how a model learns — tuning its numbers to shrink its error, over and over.
- Explain why a model predicts likely patterns rather than knowing facts, and where that makes it fail.
- Judge an AI claim, and tell which tasks suit a machine and which need a human.
Course complete
You finished every lesson. Put your name on it.
Module 1 — What a model actually is
Rules versus examples
Explain why AI learns from examples instead of following written rules.
A model is a machine that makes a guess
Explain that a model is a function with adjustable numbers that maps an input to a prediction.
Learning means tuning the numbers
Explain that training adjusts a model's numbers to shrink its error on examples.
Module 2 — How a model learns
The training loop: guess, check, nudge, repeat
Trace one training step, and explain the step-size trade-off.
Neural networks: stacking simple parts
Explain how simple units in layers learn complex patterns.
The data is the teacher
Explain why the training data sets both what a model can do and its blind spots.
Memorising versus learning
Tell overfitting from real learning, and explain the train/test split.
Module 3 — Why it behaves the way it does
It predicts, it doesn't know
Explain that a model generates likely output from patterns, not looked-up facts.
Why AI makes things up
Explain why a model can be fluent, confident, and wrong.
Bias: the mirror of the data
Explain how bias enters through the data, and why it is hard to remove.
Confidence and the black box
Read a model's confidence correctly, and explain why its decisions are hard to explain.