AI Outperforms ER Doctors in Diagnostic Cases, Study Points to Collaborative Care
- On Thursday, OpenAI's o1-preview model diagnosed real-life medical scenarios as well as or better than physicians in a study published in Science, marking a significant advance in clinical AI reasoning capabilities.
- Harvard Medical School and Beth Israel Deaconess Medical Center researchers tested o1-preview against two attending physicians using 76 real emergency department cases from a Boston hospital, where the AI handled uncertainty more effectively than previous models.
- In triage cases, o1-preview achieved 67.1% diagnostic accuracy compared to 55% and 50% for the attending physicians; the model suggested a helpful diagnosis in 97.9% of cases analyzed by researchers.
- Dr. Adam Rodman of Beth Israel Deaconess Medical Center cautioned the findings do not support replacing doctors, urging rigorous prospective clinical trials before integrating such tools into practice.
- Future clinical use requires assessing safety and equity, as current tests rely on text-only inputs and fail to capture visual and auditory signals clinicians regularly interpret when treating patients.
23 Articles
23 Articles
AI models now outperforming doctors at diagnosing patients in ER
According to an NPR report, researchers found that an advanced AI reasoning model from OpenAI significantly outperformed experienced physicians in diagnosing patients and guiding care decisions, using real-world emergency department data. In a compelling example from the study, a patient arrived at the hospital with a pulmonary embolism. After initial improvement, symptoms worsened, leading the […]
Paging Dr. AI to the ER? Artificial intelligence shows promise as tool for emergency room diagnosis
As hospitals increase their use of artificial intelligence to improve patient care, a U.S. study has found that an AI tool has proven effective at making emergency medical diagnoses — even better than actual doctors, in some cases.
New study suggests AI is starting to outperform doctors, sheds light on growing capabilities
One of the first studies conducted on AI's ability to perform complex medical reasoning tasks has been published in "Science" magazine on Thursday. It has found that AI is often performing better than human physicians in complex medical diagnosis and reasoning.
Coverage Details
Bias Distribution
- 61% of the sources lean Left
Factuality
To view factuality data please Upgrade to Premium

















