Glossary

Artificial Intelligence, software that learns from examples instead of only following fixed rules.

The trained program that makes predictions after learning from data.

The examples (images, text, numbers) an AI learns from.

A step-by-step recipe a computer follows to solve a problem.

The answer a model gives for a new example.

A repeating clue in data that a model uses to make guesses.

A way of building AI: show the computer many examples and let it figure out the rules.

The process of teaching a model by showing it many examples.

The correct answer attached to an example so the model can learn.

A measurable clue (color, shape, size) a model uses to tell things apart.

A model loosely inspired by the brain, with layers that learn features step by step.

An organized collection of examples used for training or testing.

Trying a model on new examples it never saw during training.

When a model memorizes training examples and then fails on new ones.

The instruction you give an AI. Clear prompts get clearer answers.

A program that answers in conversation, powered by a language model.

An AI trained on lots of text to predict which words come next, so it can write and answer.

A chunk of text (often part of a word) that a language model reads and predicts.

When an AI confidently states something false, like inventing a book that does not exist.

Improving step by step: read the AI's answer, fix your prompt, and try again.

AI that understands images by learning patterns of edges, colors, and shapes.

The tiny colored dots that make up an image; computers see pictures as grids of them.

A model that sorts examples into categories, like spam or not spam.

AI that creates new content, text, images, sounds, instead of just sorting.

AI-made photos, videos, or voices of real people that look genuine but are fake.

Unfairness that happens when training data leaves people or things out.

Building and checking AI so it treats different people and groups equally well.

Controlling who sees your personal information. Share as little as possible with AI tools.

Permission given freely before using someone's data, photo, or voice.

Rules and filters around an AI that block harmful stuff and keep it on track.

Where information comes from. Reliable sources say who wrote it and how they know.

A machine that senses and acts in the world. AI can be its brain, but many robots only follow fixed rules.

A device that measures the world (light, sound, motion) and turns it into data for computers.

An AI suggestion based on patterns in what you and others liked before.

AI that converts spoken words into text, the first step when you talk to an assistant.

AI that rewrites text in another language by learning from millions of translated sentences.