ACR keynote: AI in medicine ‘could be exciting or it could be terrifying’
Click Here to Manage Email Alerts
SAN DIEGO — If wielded appropriately, artificial intelligence could transform health care for the better by improving diagnoses and allowing physicians to provide more individualized care, said the keynote speaker at ACR Convergence 2023.
He also said it “could be terrifying.”
“In medicine in particular, the excitement and anxiety about AI have accelerated in the last year,” Avi Goldfarb, PhD, Rotman Chair in Artificial Intelligence and Healthcare and professor of marketing at the Rotman School of Management, at the University of Toronto, told attendees.
Goldfarb noted that the launch of ChatGPT and other such large-language models have been heralded by some as having the ability to “change the course of history,” a development which some may find “terrifying or fascinating,” he said.
The “fascinating” part pertains to whether AI will provide revolutionary improvements in patient care. The “terrifying” part, meanwhile, relates to the question of whether it will replace the human element in medicine.
“The optimistic view is that we are on the verge of the ‘Jetsons’ or the robots from ‘Star Wars,’” Goldfarb said. “They can do everything we do and they listen to us.”
However, the flip side of this hypothetical is the “Terminator” or “The Matrix,” where the robots are just as functional but do not listen to humans, he added.
For Goldfarb, the reality, at the moment, is that AI and machine learning are simply helping to make predictions. In medicine, this means diagnosis.
“Prediction has gotten better, faster and cheaper,” he said.
However, Goldfarb assured attendees that humans will always be necessary in the predictive component of patient diagnosis.
“You take data about a patient’s symptoms and fill in the missing information about the cause of those symptoms,” he said.
The larger concern, for Goldfarb, is that of all the industries that are using AI technology to make advances, health care is “way at the bottom,” he said. “When you look at AI applications in health care, it is not looking good.”
Goldfarb encouraged attendees to embrace the technology to improve the lives of patients.
“If we have machines that can do diagnosis as well or better than 90% of medical professionals, then we can think about a transformation of the medical industry,” he said.
According to Goldfarb, allowing nurses, advanced practice practitioners and other providers to oversee the gathering of symptom data and then input that data into AI or machine learning programs could save rheumatologists time, and allow them to offer more personalized care in the clinic. Moreover, these types of technology could be used by health care professionals anywhere, allowing them to provide care to rural and underserved populations.
Such a transformation may be necessary given the workforce shortage that is coming in rheumatology, Goldfarb added.
“This is a very exciting vision of what could happen in a particular industry,” he said. “It could be exciting or it could be terrifying.”