Systems that produce rather than classify.
THE DEFINITION
AI that creates new content: text, images, audio, code, video.
HOW IT DIFFERS
Traditional machine learning classifies or predicts. Generative AI produces.
HOW IT WORKS BROADLY
It learned patterns from large amounts of existing content and generates new content following those patterns.
WHAT IT PRODUCES
Novel output rather than a selection from a set. Each generation is new.
WHAT IT IS NOT DOING
Copying its training data, usually. It learned patterns, not passages.
Though models can reproduce memorised content that appeared many times.
THE PRACTICAL IMPLICATION
It produces plausible content. Plausible is not correct.
WHERE IT IS MOST USEFUL
First drafts, variations, and structured transformation of material you provide.
RELATED TERMS
Large language model, hallucination, prompt.
