Term explained
What is generative AI?
Generative AI describes systems that produce new content — text, images, audio, video or code — by predicting what plausibly comes next, based on patterns learned from large amounts of data.
In plain English
Most earlier AI classified things: is this email spam, is this a dog, will this customer churn? Generative AI creates things. Give it a prompt and it produces text, a picture, a piece of music, a video clip or working code.
Underneath, these systems are doing something simpler than it looks: predicting, over and over, what element plausibly comes next or what a noisy image should look like when cleaned up. Because they learned from vast collections of human-made material, their predictions tend to look like something a person might have produced.
Why it matters
Generative AI has changed who can produce certain kinds of work. Drafting, summarising, translating, illustrating and prototyping — tasks that used to require a specialist or several hours — can now be started in seconds. That shifts the bottleneck from producing a first draft to judging, editing and verifying it. It also raises live questions about copyright in training material, synthetic media used to deceive, and the value of creative labour.
An example
A small charity needs a newsletter. A staff member types a few bullet points about recent events and asks a generative tool for a friendly 300-word write-up. The tool returns a serviceable draft in seconds, which the staff member then corrects, cuts and gives a human voice before sending.
What to watch out for
Fluency is not accuracy. These systems optimise for output that looks right, so they will invent plausible facts, citations, quotes and statistics with complete confidence — often called hallucination. They also reflect the style and the biases of their source material, and they cannot reliably tell you where a particular claim came from. Treat the output as a draft from an enthusiastic but unverified assistant, never as a source.