The basics
Generative AI is artificial intelligence that creates new content, such as text, images, music, code, or video, from patterns it learned in existing examples.
Most older AI systems judged things: is this email spam, is this a cat, will this customer leave. Generative AI makes things instead. You describe what you want in plain words and it produces a fresh essay, picture, or tune. The output is new, not copied from a file, though it is shaped by everything the model saw during training.
Under the hood, a generative model has learned what typical examples look like. A text model has learned which words tend to follow which. An image model has learned which shapes and textures usually go together. When you ask for something, it uses those patterns to build a likely answer piece by piece.
Because it works from likelihood rather than facts, generative AI can produce confident mistakes. It is superb at drafts, brainstorming, and style, and much weaker at exact facts or fresh news. Treat its output as a first draft that needs checking, not as a source.
You type a few sentences about your grandmother's garden, and a generative model writes a rhyming birthday poem and suggests a melody to go with it.