What’s Left for Me?
Source transcript: What’s Left for Me? - Raw Transcript
I’m generally positive about artificial intelligence. More than positive, really. I find it exciting. But every once in a while I try to look a little farther ahead and think through where this might actually be going, and when I do that I eventually get to a slightly uncomfortable question: what’s left for me?
I really like my job. I get to see people all day, which is fun. They tell me about their lives, which is also fun. Sometimes we are talking about a specific medical problem and sometimes we are mostly just talking about life. I get to think through puzzles, come up with solutions, try to make sure someone is healthy, and help them navigate a healthcare system that is often harder to navigate than it should be.
The strange thing is that I am probably not the best in the world at any individual part of it. There are certainly better diagnosticians than me. There are probably doctors who are better at talking with patients, better at building relationships, better at navigating systems, better at coaxing out the thing someone is reluctant to say, or better at getting someone to change a behavior that would improve their health. There are a lot of people in the world, so it would actually be pretty surprising if I happened to be the best at any one of these things. I suppose I am, by definition, the best at being me, but I am not entirely sure what that means. It sounds like the title of a children’s book.
What I have historically been is reasonably good at a lot of these things at the same time, and that turns out to be useful. Medicine bundles together a strange collection of abilities. You have to gather information, decide which information matters, recognize patterns, understand biology, understand people, communicate uncertainty, build trust, navigate institutions, and occasionally convince another human being to do something they really do not want to do. There have always been people better than me at each individual component, but you generally could not put all of those people in the room with the patient at once.
Artificial intelligence changes that. The common line about AI is that this is as bad as it will ever be, and directionally that seems right. Whatever these systems can do today, we should probably assume they will do considerably better later. And one of the places where that matters most is something as basic as asking the right questions.
A patient comes in wondering what is wrong, and often the hard part is not knowing the answer. The hard part is figuring out what question we are actually trying to answer. That depends on getting the real story, and getting the real story depends on honesty. People are not always honest with doctors, especially when embarrassment or shame is involved. Sex, drugs, alcohol, money, relationships, weight, adherence, all of these things can change what someone is willing to say. I tell patients all the time that they should just tell me the truth because I am not there to judge them, but perhaps people are actually most honest when they are sitting alone at home talking to a computer. Although actually I am not sure, which is a strange truth of 2026. We all use computers all the time, and I have no idea what people do with their computers. But it seems plausible that talking to something that is not alive removes at least some inhibition. There is no raised eyebrow, no awkward pause, no fear that the person sitting across from you will think differently about you afterward. Perhaps eventually we will anthropomorphize these systems enough that this advantage disappears and people will become embarrassed in front of an AI too. Humans are weird.
But if AI becomes better at asking questions, better at listening, better at noticing contradictions, better at understanding medicine, better at remembering everything, better at explaining things, and better at tailoring its approach to the individual person sitting in front of it, then a lot of the things that currently make someone a good physician begin to get bundled into one system. That is the part that leaves me wondering where exactly I will fit in.
There are plenty of easy answers. Medicine is heavily regulated. Physicians have licenses. Someone has to sign things. Someone has to take responsibility. AI cannot currently walk into a room and drain an abscess, reduce a fracture, or freeze a wart. So perhaps physicians remain necessary for quite a while. But I do not find that especially reassuring. I do not want my professional value to come from regulation. I do not want the answer to be that I remain useful because the law requires someone with my credentials to stand nearby. I want to actually be useful.
Unfortunately, a lot of the physical things that remain obviously difficult for AI are not the parts of medicine I particularly love. I have never gotten out of bed excited because someday I might get to freeze a wart or drain an abscess. Why that is will probably be good material for future thinking. I think the stress of physically changing someone with my own hands, however small the intervention, is not something I particularly enjoy. But I digress.
As far how how to deal with the existential threat of AI, the answer I have arrived at, at least for now, is not to spend much time trying to defend myself against the technology. It is to get better at using it. If my patients are going to have access to an extraordinarily capable intelligence, then part of my job becomes figuring out how to use that same intelligence to take better care of them. Perhaps the physician of the future is not the person who knows more than the machine. That seems like a losing competition. Perhaps the physician is the person who becomes unusually good at working with it.
That may mean knowing when to trust it and when not to. It may mean knowing what questions to ask, how to frame a problem, and how to combine its enormous breadth with the weird, incomplete, contradictory reality of one particular human being. At least, that is the bet I am making. Turning my back on these tools because they threaten parts of my profession seems like turning my back on the future.
Still, I suspect some version of this question is going to become increasingly common. Where do I fit? Not just for doctors, but for writers, programmers, lawyers, teachers, designers, scientists, accountants, and probably eventually nearly everyone whose identity is tied partly to being good at something. We have spent our lives accumulating abilities. Then something appears that can accumulate abilities much faster than we can.
That does not necessarily make the future bleak. Standing at the edge of something enormous can produce two very different sensations. You can look down and see the fall, or you can look up and see the heavens opening above you. We are entering something genuinely new, and we do not really know how the physics work after we jump. For now, I am hanging around the edge. I am not depressed about it. Mostly I am excited, and trying to figure out what I should become next.