Generative AI

Older AI mostly "recognized" things — filtering spam emails, spotting faces in photos. Generative AI works the opposite way: it "makes" things that didn't exist before — text, images, music, video, even code.

To put it simply, if older AI was "a judge who looks at a picture and guesses what it shows," generative AI is "a painter who takes your order and paints the picture for you." Almost everything in the AI boom since ChatGPT launched in 2022 is about this "creating" AI.

Where the generated output comes from is often controversial. Because generative AI learns patterns from massive amounts of human-made text and images, questions about copyright, jobs, and fake content always follow it. When you see these debates in the news, this is the background.

"The generative AI market is projected to grow to $XX billion next year" — this term groups together AI services that create content, like ChatGPT, Claude, Midjourney, and Sora.

1. Sign up for free at ChatGPT (chatgpt.com) or Claude (claude.ai).
2. Try typing: "Write a polite 5-sentence opening greeting for a new neighborhood cafe."
3. If you don't like the result, follow up with "make it shorter" or "make it more upbeat." This back-and-forth is basically all there is to using generative AI.