Researchers from Harvard Medical School and Beth Israel Deaconess Medical Center have discovered that artificial intelligence has surpassed doctors in clinical reasoning tasks. Further studies are needed to assess the real impact of AI in practice.

Researchers from Harvard Medical School and Beth Israel Deaconess Medical Center conducted a comparative study between artificial intelligence and human medicine in clinical reasoning tasks. The results show that large language models (LLMs) outperformed doctors in decision-making, diagnosis identification, and case management. Arjun Manrai, co-author of the study, emphasized that although AI has demonstrated superior results, its implementation in medical care requires rigorous studies to assess the impact. The researchers used OpenAI's model o1, which excelled in most tests, especially in situations with limited information. However, experts warn that the use of AI in clinical decisions must be carefully monitored, as unnecessary test suggestions may arise. The study has limitations, focusing on a preliminary version of the model, and the researchers call for further studies to better understand the collaboration between humans and AI in the medical field.