Canadian researchers from Western University have determined from studying the results from 150 test cases, that the ChatGPT AI diagnosed medical conditions with less-than 50% accuracy. Online publication TechXplore reports that a team led by Professor Dr. Amrit Kirpalani "...found that ChatGPT was only 49% accurate when it came to providing the right diagnosis."
The team used a recognized method used for testing medical practitioners called Medscape Clinical Challenges. From the TechXplore article: "The researchers found it struggled with interpreting test results and sometimes overlooked critical information that was relevant to the diagnosis. "
