Learning from Examples

An example is something you show AI so it can learn. If you want AI to know what a cat looks like, you show it lots of cat pictures.

If you only show AI pictures of red apples, it might think green apples are not apples. Showing many different examples helps AI learn better.

AI usually needs thousands or millions of examples to learn well. The more complex the task, the more examples it needs.

AI does not learn everything at once. It gets a little better with each batch of examples, slowly improving over time.

People decide what examples to give AI. This is an important job because the examples shape what AI learns and how it behaves.

If examples are unfair — like only showing one type of person — AI might treat others unfairly. Fair examples help AI be fair to everyone.

AI can learn from sounds too. By hearing thousands of different songs, AI learns what music sounds like and can suggest songs you might enjoy.

Every time you use an app or click on something, you might be giving AI an example. Your choices help AI learn what people like.

We only use examples that are safe and appropriate. Grown-ups make sure AI learns from good, helpful examples — not mean or scary ones.

After AI learns from examples, people check its work. If something is wrong, they give it more examples to fix the problem.

AI can keep learning from new examples over time. As the world changes, new examples help AI stay up to date.

You can teach AI by showing it many pictures. If you show it 1000 pictures of apples, it learns the shape, color, and look of apples.

AI learns by studying many examples — pictures, words, and sounds. People choose fair, safe examples and check AI's work. More data means better learning.

AI can also learn from words and sentences. By reading millions of sentences, AI learns how words fit together and what they mean.

Just like you get better at riding a bike with practice, AI gets better at tasks when it practices with more examples.

AI can learn to sort things into groups. It can learn to tell cats from dogs, or sort shapes by size and color after seeing examples.

AI learns from both right and wrong examples. Showing it what a cat is AND what is not a cat helps it learn the difference.

When we put labels on examples — like 'this is a dog' — it helps AI know what it is looking at. Labels are like name tags for data.

In some games, AI watches how you play and learns. If you are really good at jumping puzzles, the AI might make harder jumps next time.

AI sometimes makes mistakes, especially when it is still learning. If it mistakes a muffin for a dog, that is okay — it needs more examples.