Skip to content

AI Glossary · Concepts

What is Generative AI?

Generative AI

Generative AI is AI that produces new content, such as text, images, audio, video or code, rather than only classifying or scoring what it is given. Large language models are the best-known kind.

Also known as: GenAI, gen AI

· Updated · Chain of Thought

Model ArchitectureMultimodal AI

  • 5 explainers
  • 8 episodes
  • 2 sources

Much of the AI in business before 2022 was predictive: a model looked at an input and gave back a label or a number, such as fraud or not fraud, or a demand forecast. Generative AI produces something new instead: a paragraph, an image, a song, a block of code. Official definitions say the same thing more formally. NIST’s 2024 generative AI profile quotes the definition from Executive Order 14110, since revoked: models that emulate the structure and characteristics of their input data to generate derived synthetic content, including images, video, audio and text. Large language models generate text, a task long studied as natural language generation; image, audio and video models, including diffusion models and earlier generative adversarial networks, generate media; and multimodal models handle several of these at once.

Generative AI and agentic AI are related but not the same. A generative model produces content when asked. An agent uses a generative model to plan and take actions toward a goal, calling tools and checking results along the way. Agentic vs. non-agentic AI covers the difference.

Consider three jobs in a support workflow. A classifier labels a ticket “billing.” A search system retrieves the relevant refund policy. A generative model drafts a reply from the ticket and that policy. The three can work together, but a fluent reply is not proof that the right policy was retrieved or applied, and a drafted refund email does not itself issue a refund; any action runs through software with its own permissions.

See AI vocabulary in plain English for this term alongside the other language of models, data, agents, and infrastructure.

Sources

Go deeper

From the conversation