Editor’s Note: For 16 years, I've published weekly essays here on Blogspot, which will continue. I’ve now begun publishing my work on a new blogging platform, Substack, and I hope you’ll join me there. Please enter your email address at this link to receive my posts directly to your inbox.
My readers know of my trepidation and angst regarding the unstoppable artificial intelligence (AI) tsunami that will disrupt our world. It will make the arrival of the internet seem quaint by comparison. Just as we cannot fully contemplate the potential benefits, we cannot fathom the potential harm and destruction it will wreak. There will AI actions that will cause harm and benefit simultaneously. For instance, once perfected, driverless vehicles will operate more safely than human controlled cars and trucks. Former drivers will now be passengers who are free to engage in professional or leisure activities. However, if you currently earn your living as a driver, then this technological milestone will a millstone. There will be many mixed results like this - winners and losers even when AI is being utilized in a legal and ethical manner.
Some AI actions will be all upside such as expediting the
discovery of more effective medical treatments or cures.
And, of course, there will be evil outcomes by design. No examples necessary for this.
As a physician, AI is already permeating my world, but I
have resisted its encroachment thus far. I have always had a wary eye towards
technology, which I believe has exacted a portion of our humanity as a cost for
its use. In one example, consider the harms suffered by teenagers
and others from social media.
For me, at this stage of my career, I am largely satisfied with the tools that I have been comfortably using. But I expect that in short order, every aspect of a physician’s professional life will be infused with AI, from the office visit to the operating room and everything in between. We already have AI technology that can be incorporated into the colonoscopy experience, which my GI group has not yet adopted. AI will identify polyps during the colonoscopy journey, highlighting the lesions for the endoscopist who will then remove them. (In time, there will be reliable technology that will be able to identify the type of polyp discovered without needing to perform a traditional biopsy. This would allow the gastroenterologist to leave polyps behind that AI has determined are entirely innocent.)
Here is a predictable consequence of polyp-identifying
AI. Physicians who rely upon AI will
lose their own polyp-finding skills. This skill atrophy will not be restricted to
the medical profession but will affect all of us in ways that we can imagine an
in ways that are beyond our reach. Do
you think that artists, musicians, chefs, composers, scientists, designers, writers,
journalists, tradesmen, lawyers, educators and the rest of us will
maintain our skills as AI moves in?
(Hint: They won't!) For physicians, AI will extend far beyond
medical procedures into all aspects of medical practice. I expect that it will be used to obtain a
medical history, perform a physical exam and to make diagnoses more accurately than most human physicians are doing today.
Those who dispute my conjecture probably would have argued 10 years ago
that telemedicine could never replace an in-person office visit.
So, here’s a few philosophical questions I pose to readers. If a machine can perform a task
better than we can, should we care if we humans can no longer compete? Should we just step aside? What will be left for us to do?
You are making an excellent point, there are so many ways for a completely informed AI bot to preform the duties of a physician, so seamlessly, with a vast array of information that it would be impossible for a physician to do as well.
ReplyDeleteI am a retired psychiatrist and psychopharmacology was my wheelhouse but when I needed help with a complex medical problem I was experiencing myself I turned to a bot. It was a private chat over almost 2 hours that walked me through my points of decision, gave me helpful new statistics and also offered a kind and sensitive ear. I even allowed myself to cry when I felt so supported in the chat and knew no physician I currently see has come close to the comprehensive care and support I received. Why? Because real physicians I see in my PPO system act as though they have no time, they are curt, impersonal and seem not to really care about my welfare. They are overworked, worried about their own lives, etc. My AI bot recognized the subtleties of what I was describing in my symptoms, history and my own personal strengths and vulnerabilities were reflected in his comments. I did work with my human physician today and of course did not mention the bot experience but her comments were shorter, less nuanced and less helpful to me as a person than the bot. She acquiesced to my suggestions about my own care and was comparatively distant. Offered appropriate changes but without the care and support of the bot. I find this so sad and yet this is our future. We have given our bots all they need to be our virtual everything, including our physicians. If the bot could have prescribed my med changes, as he carefully suggested lifestyle changes I would have preferred it. I'm sorry and sad and yet, that's where we are.
@anonymous, thanks for your personal and thoughtful comments. I suspect you will find this recent article of interest. https://www.nytimes.com/2025/11/16/well/ai-chatbot-doctors-health-care-advice.html?smid=url-share
ReplyDeleteTelemedicine can never replace an office visit for obvious reasons - you need a hands-on person with a stethoscope listening to your heart, lungs & carotids. You need a human being feeling your abdomen. And a machine can't interpret nuances in facial expression and therefore can't " read" a patient. Let's get back to practicing medicine the way it was meant to be.
ReplyDeleteThe important question is not whether physicians skills will deteriorate with AI. Of course they will in some areas just as I am no longer good at long division once I had a calculator and never could do square roots. The important question is whether patient morbidity and mortality will increase or decrease with widespread physician use of AI. We don't know the answer to that question right now and it needs to be studied. Your example of driverless vehicles leading to less accidents to the detriment of paid drivers is a good one. If the analogy is less patient morbidity at the expense of physician skills in some areas, I would take that outcome just as I take the benefit of driverless vehicles now.
ReplyDeleteIf AI takes over, skills will deteriorate. However, as a tool of instant, on demand, up-to-date medical knowledge it makes me a far better doctor at least as a generalist. Rather than either assuming my currently held sense of my knowledge on a topic is solid or engaging in a very sketchy search of the internet or out date textbooks, I'm getting surprisingly good data on actual clinical situations throughout the day and as issues arise, which allows better retention. Rather than or in addition to formal blocks of general CME, AI provides constant, up to date, on demand CME. Especially a bit removed from my last general boards or residency, I am fast becoming more informed than before AI tools. There is no doubt a double edge to this sword, but one edge is the one I'm seeing.
ReplyDelete@Don, if I were to have asked you 15 years if you thought that a driverless car would be safer than a human operated one, you might have dismissed that outright. Think of all of the technological achievements that we currently enjoy that we would not have believed could be possible years ago. Such will happen, I predict, in AI medicine.
ReplyDelete@anonymous, appreciate your comment and your optimism. The other 'edge' is lurking just beyond our view, I fear.
ReplyDelete