ARTIFICIAL INTELLIGENCE – SURGEON’S HEIRLOOM OR AN ENTRUSTED PENALTY?
Dr. Geeta Sundar, Department of Neurosurgery
Recently, a friend of mine asked me if I would be a redundant entity once AI bloomed in the way Elon Musk was voicing its spectrum. Surely not, I was baffled that someone could consider that a doctor’s role would be reduced to nothing in the wake of an AI apocalypse. But in all reality, once time had done its bit, I realized that maybe I had reacted too soon.
Yes, AI has brought about a lot of changes in the world of a surgeon, more so into a neurosurgeon’s world. It has revolutionized incisions, precisions, surgical techniques, navigation guided biopsies and tissue sampling, allowed for better healing, reduced complications, hospital stay; and also played a major role in publications and research writing.
But I am reluctant to allow it to enhance. Reason? Many. Personal and professional.
Professional – AI can develop on what it is fed, on what information it receives. An educated patient of mine, last week came to the outpatient care with a self-made diagnosis. He was well read to know that something was wrong with him, smart enough to get an imaging done, and get it reported, and instead of Google, he used ChatGPT to understand his diagnosis. But he took a shop-shoddy image of his MRI and uploaded it into ChatGPT – that diagnosed him to have a pituitary lesion instead of a pineal gland pathology. And cut back to when he walks in, talking to me about hormones, and his visual status, asking me about an endoscopic trans nasal surgery; and when I appear confused about his findings on the MRI film I hold up, I see a posterior third ventricular lesion at the pineal region – he is unkind to comment that I don’t know what I am talking about. #Pssh.
AI can only assist us; it would be futile to give it godly power as Marvels Cinematic Universe would otherwise enable J.A.R.V.I.S. or Vision to have. This concept is sometimes forgotten in the routine that we follow. Once we know, and taste the ability of AI, it trickles into our lives as a means of a savior. But it takes time with AI – as it does with everything. Not all aspects of life can be sorted by AI. I had a recent case presentation of a brain lesion, that was very vague, not well defined, not matching on the various MRI sequences to anything that could help me pinpoint a diagnosis. I was at a wit’s end, been through many papers, my exhaustion into understanding the pathology and preparing for the presentation took the better of my otherwise too proud self, and I looked it up on ChatGPT. What a fool! The diagnosis left me nonetheless too pleased – “it can be glioma, gliosarcoma, or other differentials like metastasis, lymphoma, etc.” Oops. This, I already knew. ChatGPT ain’t needed for this one liner, man! What I really needed was a run-down of the various differentials and why it could be one, versus not the other. Only a person with enough experience of having seen thousands of such lesions could be able to differentiate. #Better ask my teacher.
AI is not the ultimate answer to everything, it is but a means to accomplish somethings easily. Earlier days, when I was an intern, Google was king. When anyone asked a hard question during class or rounds, with the great ability to Google, with our eyes focused elsewhere, we’d have answers skimmed off the webpages. But today, its worse. Not just hard answers, but the generation these days looks up everything on ChatGPT. My juniors are keener to pay a subscription for ChatGPT than pay their cafeteria bill. Ask them hyponatremia correction, they ChatGPT the formula and calculation. Ask them the amount of fluid correction, dosing, SIADH vs Cerebral salt wasting treatment, steps of surgery, latest clinical trial – ChatGPT is all they know. Not even the standard Bible textbooks that have been recommended and passed down to us from strong abled, stalwarts of all times, have been touched or opened. Youman’s, Greenberg’s textbook lies hidden in some corner, completely undeserving of attention. In comparison to our teachers and their gurus, who had to scout for information across books, and libraries, it’s all now available at a text. Is it really worth it? Have we made it too simple for us and far lazier for the upcoming generations? #Very sad.
Personal reasons? I am not keen to allow AI to take over my life (we already allow Alexa to know our favorite songs, lock the doors, allow Tesla to drive us to places), I would like to maintain some individuality, thank you very much, and represent that little 5% that differentiate my genome from yours. I would also like to very much challenge myself and push myself to adapt to a system where there are hardships and I have to work an appetite for those challenges and accomplish them on my own with least help, remember the “survival of the fittest”; otherwise, what am I teaching and leaving behind for the coming next generation of young neurosurgeons? I would like to learn the hard way, brace through the harder pathway before I fall into an easy pathway with AI, and never know how to get myself out of a surgical fix. We never run the day we are born, do we? We learn neck control, crawling, walking and then running, all in the proper course of evolution and growing, and I think personally as well, it should be the way we use AI.
Truth be said, in all due where its worth, AI has been a boon. It has made our life simpler, easier, but it resorts a learning curve and many of us have had to change with it, and learn to adapt it into our lives. I remember my first interaction with BrainLab® in 2016 in NYC, and till then I hadn’t even known what neuronavigation could assist with. Walk forward and today we are doing DTI assisted surgeries, 3D reconstruction with approaches, minimally invasive endoscopic surgeries, DBS, frameless stereotaxy, and the major breakthroughs have been with Gamma Knife and Stereotactic surgeries, Proton therapy, soon to implode into many more aspects in the coming decade that will help us treat patients. In coming years, I hope we can use AI like how the Matrix movies did it, and to read thoughts, understand integration of neural signals better, and comprehend memory, speech, neurodegenerative diseases, tumors and their natural history.
But that doesn’t mean, we bow out to this entity. It doesn’t mean we let it run amok. AI could never replace a surgeon’s/doctor’s compassion or empathy. AI is also bound to fail, and have setbacks, machinery issues. There are instances in news, where people are now finding AI assisted dolls/personifications to fill up their lives and represent companionship. The day we fall to this, to losing out on human touch, to losing out on our humanness, our ability to socialize, we will befall a huge calamity – I fear, I am skeptical that could represent a void that once filled with AI, we may never come back to what our society today is. A collaboration between us and AI is doable, sure, but with a back-off-kill switch, if we ever cross a line. #God forbid.
All images attached – source from Google®
