Politico July 3, 2024
DIAGNOSIS
Advanced AI chatbots performed as well as doctors by some measures in a recent study evaluating diagnostic abilities.
Researchers from three Boston hospitals, Mass General Brigham, Beth Israel Deaconess Medical Center and Brigham and Women’s, compared diagnoses — and the reasoning behind them — made by doctors and GPT-4, the bot from San Francisco’s OpenAI, for 20 cases.
GPT-4 and doctors came to the correct diagnosis at similar rates, but the humans made fewer mistakes in their reasoning to get to those diagnoses, according to a new study.
One author predicts doctors will eventually use bots to help them make diagnoses but doesn’t expect them to replace doctors.
Even so: The authors acknowledged that the evaluation criteria favored the...
Today's Sponsors
Today's Sponsor