The wide-ranging possibilities and “black box” nature of AI raise the danger of over-exuberance in some people and fear in others. We need a little of both. Especially in the early days, AI will be no substitute for astute clinical judgment.
This was made clear to me last year when my grandson was born prematurely in London. One night in the neonatal intensive-care unit, an experienced nurse thought “something wasn’t right” about how his blood was clotting after his many pinprick blood tests. On a hunch, she sent off specialized laboratory tests and discovered that he had severe hemophilia. If she hadn’t pursued that intuition, the problem might not have been diagnosed for weeks or months, and he might have been in danger of severe bleeding from simple manipulations and blood tests.