GPT-4 Correctly Diagnoses 39% of Cases, According to Beth Israel Researchers

Researchers at Boston’s Beth Israel Deaconess Medical Center have found that AI tool GPT-4 can correctly diagnose challenging medical conditions 39% of the time. The tool was tested on 70 medical cases, with the results showing that it could include the correct diagnosis on its list of potential diagnoses in 64% of the cases. The findings come after GPT-4 failed an exam often taken by urologists in training. Despite this, the researchers at Beth Israel Deaconess found generative AI to be a promising adjunct to human cognition in diagnosis, given its potential to help physicians make sense of complex medical data and broaden or refine diagnostic thinking. While researchers admit that privacy issues need to be addressed, they believe these are exciting findings for the future of diagnosis and patient care.