Generative AI is AI that creates new content — text, images, audio, video, or code — rather than just analysing or classifying existing data. It’s the category that includes ChatGPT, Midjourney, and most of the AI tools that went mainstream.
What it means in plain English
For most of AI’s history, the goal was analysis: is this email spam, is this a cat, what number is this? Generative AI does something different — it produces. Give it a prompt and it generates something new that didn’t exist before: a paragraph, an illustration, a melody, a block of code. The output isn’t copied from a database; it’s newly assembled based on patterns the model learned from huge amounts of training data.
This shift from analysing to creating is what made AI suddenly feel magical to the public. A tool that tells you whether a photo contains a cat is useful; a tool that draws you any cat you can describe is a different kind of experience entirely.
A simple example
Typing “a watercolour painting of a fox in a snowy forest” into an image generator and receiving an original picture is generative AI. So is asking a chatbot to write a poem, or having a coding assistant produce a function from a description. In each case, the AI creates something new to fit your request.
Why it matters
Generative AI is the term for the current wave of AI tools reshaping creative and knowledge work. It’s built on the same foundations — machine learning, deep learning, neural networks — but points them at creation rather than classification. Understanding it as a category of capability (making new content) helps you see what unites tools as different as a chatbot and an image generator.
Related terms
- Large Language Model — generative AI for text.
- Diffusion Model — a common technique behind generative AI for images.
- Prompt — the instruction you give a generative AI.
- Hallucination — a risk when generative AI invents plausible but false content.
Frequently asked questions
What is generative AI?
It is AI that creates new content — text, images, audio, video, or code — rather than only analysing or classifying existing data.
What powers generative AI?
Modern generative AI is powered by large models such as transformers (for text) and diffusion models (for images), trained on vast datasets.
