With the growing role of artificial intelligence in medicine practice and management, the question cannot be avoided: What role will human doctors and medical staff play in 21st-century medicine?
In an article in the New Yorker, Pulitzer Prize winner Siddhartha Mukherjee discusses the role of artificial intelligence in diagnosing diseases. Mukherjee first examines how doctors come to a diagnosis, then discusses the challenges facing machine diagnosis. Ultimately Murkherjee believes that machines will not replace doctors, but support them and help provide more accurate diagnoses.
Artificial Intelligence in Medicine: Knowing Knowing
An important concept highlighted in this article is the difference between “knowing that” and “knowing how”. It is easy for people, and computers for that matter, to be taught a set of rules where they can identify an object and provide information about it. It is an entirely different matter for a person or computer to learn how something is done.
How does this distinction apply to artificial intelligence in medicine? Mukherjee explains that as doctors gain experience, they move from “knowing that” to “knowing how”. They use the experiences they have to recognize patterns and ultimately become better at forming a diagnosis.
In contrast, computers have traditionally operated with sets of rules that allow them to come to some result. One could compare this to “knowing that”. But now scientists are working on diagnostic machines that can analyze an exceptional amount of data, but instead of giving the computers a set of rules, they provide the machine with a directive and a set of data that is proven to be true. The computer can then interpret that data and use it as a baseline to interpret new data. The most intriguing aspect is that this way the machine is actually learning and improving over time, in the same way a child learns to ride a bike or a physician gains and applies experience–in other words, “knowing how.”
Artificial Intelligence in Medicine: Detecting vs. Detectives
But how far can this progress go? There are clearly many applications for AI in healthcare, but when it comes to diagnosis, Mukherjee and others believe that doctors will still be needed. Machines may become more accurate than doctors at diagnosing issues, but they are not capable of explaining why and how patients have contracted their illnesses. They may also struggle to identify a patient’s possible future health problems. While technology can improve, it will not be able to replace doctors completely in clinical settings, but instead provide a powerful supporting resource to help physicians diagnose disease and care for patients. Or as Mukherjee puts it, “In the realm of medicine, perhaps the ultimate rewards come from knowing together.” This indeed might point us to the future of artificial intelligence in medicine.
Source: “A.I. Versus M.D.: What happens when diagnosis is automated?“
ArbiMed offers medical inventory management software that can lower costs and optimize spending for your practice or facility. To learn more please visit our website, or contact us today to schedule a free demo.