What AI is (and isn't)
Artificial intelligence (AI) is when computers do tasks that usually need human thinking, like recognising faces or understanding speech. Many AI systems learn from examples instead of only following rules written by people.
📌 Key points
- AI stands for artificial intelligence. "Artificial" means made by people.
- An ordinary computer program follows rules that a programmer writes, step by step. A calculator is a good example.
- Many AI systems learn from examples instead. People give the computer lots of examples and it finds the patterns itself.
- AI can do some tasks that usually need human thinking: recognising pictures, understanding speech, translating languages and suggesting what to watch.
- AI does not think or feel like a human. It has no feelings and no opinions of its own, and it does not really understand things the way we do.
- Most AI is good at one narrow job. An AI that plays chess cannot drive a car.
- Not every machine or robot uses AI. A washing machine that follows a timer is just following rules.
- People design, build and check AI. People are responsible for how it is used.
📖 Key words
- Artificial intelligence (AI)
- Computer systems that can do tasks that usually need human thinking, such as recognising pictures or speech.
- Program
- A set of instructions that tells a computer what to do.
- Rule
- An instruction such as "if the room is colder than 18 °C, turn the heater on".
- Algorithm
- A set of steps to follow to solve a problem or do a task.
- Learning from examples
- Finding patterns in many examples so the computer can deal with new ones.
- Narrow AI
- AI that is good at one particular task only.
✏️ Worked example
Two programs sort emails. Program A uses one rule: "If the email contains the words FREE PRIZE, put it in spam." Program B was shown thousands of emails labelled "spam" or "not spam". Which one is learning from examples?
- Program A only follows a rule a person wrote. If a spammer writes "Fr3e Prize" instead, the rule misses it.
- Program B looked at thousands of labelled emails and found patterns that spam often has, such as strange links or asking for money.
- When a new email arrives, Program B compares it with the patterns it learned and makes a prediction: spam or not spam.
- Answer: Program B is learning from examples, which is how most modern AI works. Program A just follows a fixed rule.
Remember: rules are written by people, but machine learning finds its own patterns from examples. Ask yourself, "Did it learn, or was it told exactly what to do?"
Test yourself8 questions from a bank of 20. A new mix every time.
Start the quiz →