Artificial intelligence learns from examples — Computing, 7–10
Some computer systems find patterns in many examples and use them to make guesses. They do not think like people, and their answers can be wrong or unfair.
Idea
In one kind of artificial intelligence, a computer studies many examples instead of receiving every rule by hand. It looks for patterns and uses them to make a prediction about a new example. If it sees many labelled pictures of cats and dogs, it may learn clues that help it guess what a new picture shows.
Why it exists
Some jobs involve too many messy examples for a person to write every possible rule. A system can compare thousands of cases and find useful regularities, such as likely speech words or unusual activity in a picture. This approach came from the problem of handling complicated patterns, not from giving a computer a human mind.
Worked example
Imagine a system judging whether a fruit is an apple. Step 1: it studies 100 labelled fruits, including colour, roundness and weight. Step 2: it notices that many apples are round and red, but not all. Step 3: a new fruit is round, red and weighs 150 grams. Step 4: the system compares those clues and predicts “apple”. That is a guess based on patterns, not certainty.
A common trap
A reasonable mistake is believing that a confident answer must be a correct one. A system can sound certain while its examples were too few, badly labelled or unlike the new case. If it learnt mostly pictures of red apples, it may struggle with a green apple. Check the evidence instead of trusting confidence alone.
Outside school
Pattern-finding systems help sort spam, suggest videos, recognise speech and translate text. They can be useful, but people still need to choose the examples, check results and notice unfair patterns. When you use an AI tool, treat its answer as a helpful first guess that deserves checking, not as an automatic fact.
Keep exploring
Other languages
Loading MyLeoNes™…