What Is AI, Really?
Loading simulation…
flagWhat you'll discover
- arrow_forwardDistinguish traditional programming from machine learning
- arrow_forwardExplain what AI, ML and deep learning are and how they relate
- arrow_forwardRecognise where AI already appears in everyday life
- arrow_forwardGive one example of a task AI is good at and one it struggles with
Telling versus learning
For most of computing history, a program could only do exactly what a human wrote down as rules. Want a computer to spot spam email? You write rules: if it contains "FREE MONEY", if it has too many capitals, if it comes from a blocked sender… The trouble is that list never ends, and spammers adapt faster than you can write rules.
Machine learning flips the problem. Instead of giving the computer rules, you give it hundreds of examples of spam and not-spam, and let it discover the patterns itself. The computer writes its own rules — far more, and far subtler, than any human could. That single shift, from "tell the computer" to "let it learn", is the engine behind almost every AI breakthrough you have heard of.
The three nested circles
Artificial intelligence is the big umbrella — any technique that makes a machine seem to think. Inside AI sits machine learning (ML): systems that improve at a task by learning from data rather than following hand-written rules. Inside ML sits deep learning: machine learning that uses many-layered neural networks, inspired by the wiring of the brain.
So every deep learning system is machine learning, and every machine learning system is AI — but not the other way round. ChatGPT, image generators and self-driving vision are deep learning; a spam filter might be simpler ML; a chess-playing program that searches millions of moves is classic AI without much learning at all. In the simulation, drag each example into the ring where it belongs.
AI is already everywhere
You use AI dozens of times a day. When your phone unlocks by recognising your face, that is deep learning on an image. When YouTube suggests a video you actually want, that is a recommender system learning your taste. When Google Translate turns Nepali into English in an instant, when a map app predicts your arrival time, when a doctor gets an early warning from an X-ray — all AI.
But AI has real limits. It is only as good as its training data, it can be confidently wrong, and it has no common sense or true understanding. It spots statistical patterns brilliantly; it does not "know" anything in the human sense. Knowing both the power and the limits is the first step to using AI well — and the rest of this course shows you exactly how that power is built.