The practice of medicine and the limits of Artificial Intelligence

It is rather short-sighted to be critical of Artificial Intelligence (AI) -- not least because it seems that every time someone says, "Technology will never be able to do …" (you can fill in the blank here), technology goes and does it. Nevertheless, I've grown accustomed to playing the fool, and so I want to ponder whether technology will in fact be able to attend to what matters in medicine. Just to be clear, though, there are many things that I think technology can do better than the human; to me those things are rather uninteresting. At the risk of angering our future, computer overlords, I will offer three domains where I think that machines will have difficulty. When a physician enters a room, she greets the patient face-to-face. She engages with the person --that which animates the face just beyond its surface properties. She is, of course, looking at or reading the surface of face, but she is at the same time looking at something that seems to dance beneath the face. Yet what animates the face cannot really be said to be beneath the face -- it is right there on the face. Let me put this differently: when greeting a person, the physician -- indeed, anyone engaging truly with another person -- finds that which is "beneath" the surface of the face, gathering in excess on the face itself.