AI vs Medicine

In the article “Artificial intelligence and medicine: hype, hubris, and the humanity we can’t replace” that appears in this edition of the Southwest Journal of Medicine, Christopher Peterson argues that AI cannot replace the humanistic side of medicine. Why not? What is the humanistic side of medicine? Can humans replace the humanistic side of medicine? These are important questions that do not have simple answers suitable for a Tweet.

The syntax for computer decisions is: if (booleanExpression) then (action 1) else (action 2). The term booleanExpression is an expression that objectively resolves into a boolean result of True or False. By objectively, I mean following the rules of mathematics such that there is no rational debate about the result. The expression action 1 is executed if the boolean­Expression resolves to True. The expression action 2 is executed if the booleanExpression resolves to False. The computer follows mathematical rules; the computer does not make human choices.

What is a human choice? Do I want apple pie for dessert? Or vanilla ice cream? The question of what do I want for dessert does not have an objective answer. The choice will depend on my very human and individual preferences. I might choose on the basis of an objective rule such as fewest calories or lowest price. I might choose on the basis of my very individual and human perception of what gave me the most pleasure in the past. If I choose vanilla ice cream, there is no rational basis for someone to argue that I have chosen wrong. A computer cannot make this choice. A computer can be programmed to make the choice using rules provided by the programmer, but there is nothing generally human about that choice; that choice merely is a reflection of the individual programmer.

An AI can provide a physician or patient with the answer to the question: which course of action will have the greatest average life expectancy? The answer based on objective information will not necessarily be correct for this individual patient because human beings are not protons and we cannot predict the future path of a human being like we can predict the future path for a proton. More importantly, the computer cannot answer the question: which course of action is best for this individual patient? The computer can only answer the question based on the preferences provided by the programmer. Those preferences cannot be generalized to every human being.

The problem becomes even murkier when morality is considered. Most people put a very high value on self-preservation. However, on occasion, an individual chooses to sacrifice self for the benefit of family, friends, neighbors, or even complete strangers. Self-sacrifice is a human choice that cannot be reduced to a utilitarian equation. Attempts to do so will result in failures. The very popular Sci-Fi series, Terminator, has a plot based on an AI attempting to wipe out humanity in order to preserve the AI self. The very popular series of Sci-Fi novels by Isaac Asimov, Foundation, has, in its later volumes, the major theme of an AI struggling with the dilemma of making choices for humanity without any human basis to do so.

Finally, humanity does not require AI to abandon human choice in favor of machine rules. When protocols cease to be guidelines, created to assist the inexperienced to make difficult choices, and become rules that MUST be followed irrespective of exceptions and outliers, it does not matter that the rules were created by individual humans rather than an AI; the result of depriving patients of their autonomy over their own individual choices is the same result.

Keywords: Artificial intelligence, humanism, medical ethics

Article citation: Berdine G. AI vs Medicine. The Southwest Journal of Medicine 2025;13(57):74–75
From: Department of Internal Medicine, Texas Tech University Health Sciences Center, Lubbock, TX (GB)
Submitted: 9/15/2025
Accepted: 9/29/2025
Conflicts of interest: none
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.