Generative AI covers models that produce new output in response to a prompt. Text models like a large language model write and reason in language; other models create images, audio, video or code. What they generate is new in the sense that it was not copied from a single source, but it is shaped by the patterns in the data the model was trained on.
It is useful to separate generation from retrieval. A generative model writes plausible content; it does not automatically check that the content is true. That gap is why LLM hallucination exists and why techniques like Retrieval-Augmented generation and AI agent grounding are used when accuracy matters.
Related: AI Assistant, Prompt Engineering, AI Automation.