What is generative AI?

Generative AI is AI that creates new content, such as text, images, audio, video or code, when asked in plain language. Older kinds of AI mostly sorted or scored things that already existed, like a spam filter deciding whether an email is spam. Generative AI writes the email, draws the picture or produces the voice. ChatGPT, Claude and Gemini are the best-known examples.

Why it matters for a business

In current marketing, “AI” almost always means generative AI. It is the kind of AI people talk to, which is why it showed up inside businesses so quickly. It also brings a particular failure: the output is equally fluent whether or not it is true, because one process produces both. Generative AI is strong at drafts, summaries, extraction and reformatting, and needs checking wherever a wrong detail would be acted on.

Example

Let’s say a sales team uses one model to score which leads are likely to buy and another to write the first email to each. The scoring model sorts. The writing model generates. Both are AI, and only the second is generative AI.

Go deeper

AI vs. machine learning, and where LLMs fit

Related terms

• Large language model (LLM)
• Multimodal AI
• Machine learning
• AI hallucination
• Foundation model

All terms, A to Z