Answer

Fine-Tuning: Teaching Old AI New Tricks

An AI starts out knowing a little about everything — like a student who studied every subject. Fine-tuning is when you take that general AI and give it extra training on one specific topic so it gets really good at that one thing. You're not building a new AI from scratch. You're just adding a focused lesson on top of what it already knows. It's cheaper and faster than starting over, but the AI needs good examples to learn from, otherwise it picks up bad habits just as easily as good ones.

Example: A hospital takes a general AI and fine-tunes it on thousands of medical records so it can accurately summarize doctor's notes — something the original AI did poorly.