Ability of AI-powered chatbot to make accurate diagnoses in complex cases

Editor's Note

This study by researchers at Beth-Israel Deaconess Medical Center, Boston, finds that generative artificial intelligence (AI)-powered Chat-GPT 4 is a promising potential adjunct to human cognition in diagnosing complex cases.

Evaluating 70 complex and challenging patient cases in the New England Journal of Medicine’s clinicopathologic conferences (CPC), the AI model’s top diagnosis agreed with the final diagnosis in 27 (39%) cases. In 45 (64%) cases, the AI model included the final diagnosis in its differential—a list of possible conditions that could account for a patient’s symptoms, medical history, clinical findings, and laboratory or imaging results.

Future research should investigate  potential biases and diagnostic blind spots of generative AI models, the researchers note.