Medical large language model

A large language model is software that has read enormous amounts of text and learned which words tend to follow which. A medical one is fed extra material: clinical guidelines, drug information, hospital leaflets, patient questions.

Ask it “which department treats stomach pain” and it does not look up a rule book. It produces the answer that fits the pattern of its training. Hospitals usually bolt on their own file of departments, doctors and opening hours to keep it tied to facts.

The weak spot is that a wrong answer reads exactly like a right one. That is why most deployments keep a human nearby.