A Harvard study just found AI can now out-diagnose physicians in the ER: ‘We’re already at the ceiling’
7 Articles
7 Articles
Artificial intelligence systems based on neural networks (such as ChatGPT, Claude, DeepSeek or Gemini) are extraordinarily powerful, but their internal functioning remains largely a 'black box', so a group of physicists from Harvard University (United States) has tried to unravel this issue.
According to a new study, advanced artificial intelligence can not only compete with doctors, but also make better decisions in certain situations. The study was conducted by researchers from Harvard and Stanford, among others, and compared the performance of AI and practicing doctors in several clinical tests, even in real-life emergency cases.
News from HD Technology. Visit www.hd-tecnologia.com for the latest news. Artificial intelligence ceased to be an occasional tool and went on to occupy a huge place in daily life. Many people already use it to study, work, resolve doubts and organize simple tasks. Now, a recent study puts on the table a more delicate question, how useful an AI can be in the face of health problems. For years, the general recommendation was not to rely on AI mode…
Artificial intelligence in a Harvard Medical School study was more accurate in determining the diagnoses of patients in the emergency department than two doctors: the OpenAI o1 model gave a correct or close diagnosis in 67% of cases versus 55% and 50% in physicians, respectively.
Read the full article on stephaneLarue.com A study published in Science by Harvard and the Beth Israel Deaconess Medical Center reveals that OpenAI's o1 model makes the right diagnosis in 67% of emergency cases, compared to 55% and 50% for two experienced physicians. A major result, however, which calls for caution before any clinical deployment.
Coverage Details
Bias Distribution
- 100% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
