Generative AI definition
Generative AI is artificial intelligence that creates new content, such as text, images, audio or code, from a prompt. It learns patterns from large amounts of training data and produces output that resembles that data, instead of only sorting or scoring what already exists.
Generative AI covers any model whose output is new material: a paragraph, a photo-like image, a voice clip, a function in a codebase. ChatGPT made the idea mainstream when OpenAI released it in November 2022, but the techniques underneath are older. The transformer architecture that most text models still use was published by Google researchers in 2017.
How does generative AI work?
A model is trained on a huge set of examples until it can produce plausible new ones by itself. Text models predict the next token, a word or a piece of one, and repeat that until the answer is finished. Most image models use diffusion instead. They start from random noise and remove it step by step until what remains matches the prompt.
Nothing in that process checks facts. The model writes what is statistically likely, which is usually right and occasionally wrong in a very confident voice. Those wrong answers are called hallucinations.
Generative AI vs. traditional AI
Most machine learning in production before 2022 was discriminative. A spam filter labels an email. A fraud model scores a card payment. A recommender ranks products you might buy. Each of them picks between options that already exist, while a generative model produces something that wasn't there before you asked.
Where generative AI shows up in website projects
On a typical B2B site build it drafts a lot of small things: alt text, meta descriptions, placeholder copy to test a layout, a header illustration for a blog post. AI code editors write components and tests, and Webflow and Framer both have AI features built into the editor.
Someone who knows the brand and the facts still has to read the output. A generated meta description that invents a product feature is worse than a plain one, and the model won't warn you that it made something up.
Generative AI also changed who reads a website. Google's AI Overviews, ChatGPT search and Perplexity write answers using web pages as sources, so a page now has two readers: the visitor, and the model summarizing it for someone else.
Related terms
- Language modelA language model is an AI model trained to predict the next word, or part of a word, in a piece of text. Large language models (LLMs) such as GPT, Claude and Gemini use that ability to answer questions, write, translate, summarize and generate code.
- AI agentAn AI agent is software that uses a language model to work toward a goal on its own, step by step. It decides what to do next, calls tools such as search, APIs or a browser, checks the result and keeps going until the task is done or it needs a human.
- Vibe codingVibe coding is building software by describing what you want to an AI model and accepting the code it writes without reading it closely. You judge the result by running it and paste any errors back into the chat. Andrej Karpathy coined the term in February 2025.
- Agentic codingAgentic coding is software development where an AI agent does the hands-on work. It reads the codebase, edits files, runs commands and tests, and keeps iterating until the task is done. The developer sets the goal, reviews the result and decides what ships.