ChatGPT o1 surpasses doctors: OpenAI's AI outperforms ER diagnoses, according to Harvard

When it comes to diagnosing patients arriving at the emergency room, a study from just a few months ago revealed that human doctors were better than OpenAI's ChatGPT. OpenAI's o1 model outperforms doctors in a Harvard study. But technology is advancing rapidly, and OpenAI's first-ever "reasoning" model, o1, has recently proven superior to healthcare professionals. Researchers from Harvard and Beth Israel Deaconess Medical Center in Boston highlighted this fact through a study comparing AI to doctors. A higher diagnosis rate: 67.1% for AI versus 55.3% for doctors. As a result, o1 correctly diagnosed 67.1% of the 76 emergency cases presented to it, while the two human doctors achieved 55.3% and 50% respectively. Furthermore, other doctors were unable to distinguish the diagnoses made by the AI from those of their human colleagues. Arjun Manrai, co-author of the study and professor of biomedical informatics at Harvard, emphasized, "I don't think our results mean that AI will replace doctors, despite what some companies might claim." He did, however, describe the team's findings as evidence of "a truly profound technological shift that will transform medicine."