Around Christmas time every year, The BMJ publishes lighthearted feature articles and original, peer-reviewed research intended to ease physicians into the holiday season and help them escape the drudgery of practice. The 2024 collection of articles is no exception, with enticing papers like “How to transport a polar bear, and other idiosyncrasies in providing emergency medical services in the Arctic” and “Living happily ever after? The hidden health risks of Disney princesses.”
My favorite article is: “Age against the machine—susceptibility of large language models to cognitive impairment: cross sectional analysis.” This article shows that large language models (LLMs), such as chatbots, perform dismally on a standard cognitive test – the Montreal Cognitive Assessment (MoCA) – commonly used to screen patients for dementia and delirium. The chatbots are challenged by visuospatial and executive tasks, e.g., the trail-making task and the clock-drawing test. “Older” versions of chatbots, like older patients, tend to perform worse on the test. The findings challenge the assumption that artificial intelligence (AI) will soon replace human doctors.
Thank goodness! As much as I embrace the use of AI to aid in medical diagnosis and treatment, I don’t want AI to replace me. This is a concern shared by many doctors, as multiple studies have shown LLMs outscoring human physicians in a range of tasks, from answering questions simulating the neurology boards to providing higher quality and more empathetic responses to patient questions. Major advancements in the field of AI have led to a flurry of excited and fearful speculation as to whether chatbots will eventually surpass human physicians.
However, AI’s poor performance on the MoCA serves as a critical reminder of its limitations. While large language models excel in pattern recognition, data analysis, and generating human-like responses, they lack the nuanced cognitive abilities, empathy, and contextual awareness that define human intelligence. The MoCA, with components that focus on visuospatial tasks and executive functioning, underscores this gap: these skills require not just knowledge but also the ability to perceive and interact with the physical world in a way that chatbots cannot replicate.
This brings us to a fundamental truth about medicine: it is not merely a technical exercise but an art deeply rooted in human experience. The role of a physician extends far beyond diagnosing diseases or prescribing treatments. Doctors engage with patients on a deeply personal level, interpreting not just symptoms but also stories, emotions, and unspoken fears. They offer a listening ear, a comforting presence, and a shared humanity that AI cannot duplicate.
The allure of AI lies in its efficiency, consistency, and capacity to process vast amounts of information. These strengths make it an invaluable tool for physicians, enhancing diagnostic accuracy, streamlining workflows, and potentially alleviating some of the burdens that contribute to burnout. However, medicine is more than the sum of its parts, and a doctor’s role as a healer, counselor, and advocate cannot be outsourced to algorithms.
This is not to say that AI does not have a place in the future of health care. On the contrary, it should be viewed as a powerful ally, augmenting human capabilities rather than replacing them. For example, AI can assist in the early detection of diseases through image analysis, predict patient outcomes using data-driven models, and provide decision support in complex cases. But when it comes to making sense of a patient’s narrative, building trust, and addressing the subtleties of human suffering, there is no substitute for the human touch.
The findings from the MoCA study reaffirm the irreplaceable role of human doctors. The results also have significant personal meaning: it vindicates my failure in the dental boards and a potential career in dentistry.
Upon graduating college, I was not at all certain I would be accepted into medical school, so I applied to dental school as well. I took the dental boards “cold.” I was not aware that a significant portion of the exam tested visuospatial orientation, which makes perfect sense considering the job of a dentist requires understanding the relationships between structures in the mouth (be careful in there!). My 3-D skills are terrible. I think my exam scores bottomed out in the lowest percentile of test-takers.
Parenthetically, research has found that spatial ability is only mildly predictive of performance in restorative dentistry practical laboratory classes, but not of learning anatomy in general or practicing dentistry with greater expertise. Likewise, good visuospatial capabilities do not necessarily make better surgeons.
Also, consider that the initial part of every patient encounter, especially in psychiatry, is the overall impression you get while talking to the patient, which requires many visual abstraction skills. I may have limitations in my right parieto-occipital lobe, where visual-spatial functions typically reside, but it has not prevented me from doing a detailed mental status examination, which begins by making general impressions about the patient.
I find comfort in the fact that while AI may excel in some areas, it may fall short in spatial abilities and other cognitive tasks – particularly those that require the uniquely human combination of intellect, intuition, and empathy. Instead of fearing replacement, we should focus on integration, leveraging AI’s strengths to complement our own and create a health care system that is both technologically advanced and deeply humane.
After all, medicine is not just about solving problems – it’s about caring for people. And as long as that remains true, there will always be a place for doctors at the heart of health care.
Arthur Lazarus is a former Doximity Fellow, a member of the editorial board of the American Association for Physician Leadership, and an adjunct professor of psychiatry at the Lewis Katz School of Medicine at Temple University in Philadelphia, PA. He is the author of several books on narrative medicine, including Medicine on Fire: A Narrative Travelogue and Story Treasures: Medical Essays and Insights in the Narrative Tradition.
