What Is Generative AI?

Most software you've used so far in these courses works with data that already exists — sorting a list, querying a database, validating a form. Generative AI flips that: given a prompt, it produces content that didn't exist a moment ago.

Generating vs. classifying

A traditional machine learning model is usually built to answer a narrow question about existing input: is this email spam? What's in this photo? What's tomorrow's likely temperature? A generative model instead produces new output — a paragraph of text, an image, a block of code — that resembles the kind of thing it was trained on, without copying any single example directly.

The common categories

Generative AI isn't one technology; it's a handful of related approaches applied to different kinds of content:

• Text — large language models (LLMs) that write prose, answer questions, summarize, or generate code, covered starting in the next lesson.
• Images — diffusion models that turn a text description into a picture, covered in lesson 6.
• Audio — models that generate speech or music from text or a short sample.
• Code — a special case of text generation, trained heavily on source code rather than general prose.

This course focuses mainly on text (LLMs) and, briefly, images — the two categories you're most likely to actually integrate into a project.