Marschall Runge

Forum Message from the Dean

AI and the Irreplaceable Human Element

Fall 2018
Share Email Print
Text: A

Never before has there been more interest in applying technological advances from other fields to health care. Currently, there is intense interest in using advances in artificial intelligence (AI) and machine learning (ML) to improve health. 

Recently, Stanford researchers used a deep learning algorithm to identify cancerous skin lesions in medical images. An Israeli-based medical analytics company announced that it had developed an algorithm capable of detecting intracranial hemorrhages, which are often missed and contribute to nearly 6 million deaths each year. Researchers here at U-M have produced an algorithm that weighs about 10,000 factors per patient to predict which ones might be susceptible to contracting C. diff. 

These and other promising breakthroughs have led some to suggest that machines eventually may supplant doctors. It can be difficult to separate reality from hype with technological advances, as is the case today with AI/ML. They are not the first (or only) dramatic and revolutionary technologies brought to health care, but rather the latest in a long line of breakthroughs that have enabled caregivers to better diagnose and treat illness. 

Experts in AI/ML believe that, most optimistically, we are many years away from being able to rely on any computer-based approach for diagnosis or treatment decisions. In fact, at a meeting of the Blue Ridge Academic Health Group held in June of this year, a group of experts opined that the day machines take over medicine in any field may never come. 

This is not to say there is not great value in applying these advances, which have been quite helpful in manufacturing and other professions, to medicine. The transition to electronic health records and online medical information over the last few decades has also vastly increased our ability to collect and analyze patient data — so much so that care providers now have far more information than they can process on their own. 

Indeed, the consensus from the Blue Ridge Academic Health Group meeting was that there are very practical, nearer-term applications of AI/ML that could greatly benefit health care providers and patients, referred to as “augmented intelligence.” Imagine the benefit of a system that could rapidly summarize and prioritize all relevant findings from past health care encounters and could, at the same time, weed through millions of publications to give the provider the latest published findings relevant to the care of the patient in front of her. 

Perhaps this is not as far-reaching as saying AI/ML will replace radiologists and pathologists, but augmented intelligence will offer an opportunity for health care providers to understand the power of these tools — as well as their limitations. 

The human body is such a complex and dynamic entity — each one unique — that medicine will never become a mathematical problem where data can be crunched to produce the single right answer. AI will increasingly inform and improve this decision-making process — a process that must be guided by physicians. The great challenge going forward is in recognizing and nurturing this irreplaceable human element, to train doctors to work with machines without becoming too reliant on them. 

Marschall S. Runge, M.D., Ph.D. 
Dean, Medical School 
Executive Vice President for Medical Affairs, U-M 
CEO, Michigan Medicine 

Photo by Leisa Thompson Photography