Study Claims AI Can Reason Like a Doctor

Researchers were surprised by AI’s performance in medical scenarios. A study revealed that AI now matches or surpasses physicians in diagnosing medical cases. OpenAI’s o1 model was tested against GPT-4 and human doctors using unknown clinical cases. The o1 model outperformed older models and physicians, correctly diagnosing over two-thirds of emergency cases. Experts cautioned on AI’s role, emphasizing further research for clinical application. They noted input limitations and the need for human oversight in AI’s integration into healthcare practices.