Short answer: Artificial intelligence (AI) is when a computer learns patterns from lots of examples, then uses those patterns to make a guess — like recognising a face, suggesting a video or finishing a sentence.
The one-sentence version for children
“AI is a computer that learns from examples instead of being told every rule.” That's it. A child who understands this sentence understands more than most adults.
How it learns: a pizza example
- Collect examples (data): show the computer thousands of photos, some labelled “pizza” and some “not pizza”.
- Find patterns (training): it notices that pizza photos often have a round shape, melted cheese and red sauce.
- Test it: show it new photos it has never seen.
- Fix mistakes: if it thinks a red frisbee is a pizza, add more examples and train again.
AI doesn't “know” what pizza is. It has learned which patterns usually go with the word.
Why AI makes mistakes
Because it only knows the patterns in its data. If every pizza photo it saw was taken in daylight, it might struggle at night. If its examples were unfair, its decisions can be unfair too — that's called bias. Chatbots can also “hallucinate”: they predict words that sound right, which isn't the same as checking facts.
Where children already meet AI
- Video and music recommendations
- Face unlock and photo albums that group faces
- Voice assistants and auto-captions
- Spam filters, maps and translation apps
- Chatbots and homework helpers
Try this tonight (10 minutes)
Play “Teach the robot”. One person is the robot and knows nothing. Others give it examples: “This is a fruit: apple. This is not a fruit: carrot…”. After five examples, test the robot on something new — a tomato! Talk about why the robot got confused. You've just explained training data, testing and edge cases.
Want more? Try the How AI works quiz together, or ask Orvi “What is machine learning?”.
Our editorial promise: we cite primary sources where possible, avoid fear-based claims and correct mistakes quickly. Spotted something out of date? Tell us.
