A recent series of experiments led by researchers at the Beth Israel Deaconess Medical Center and Harvard Medical School found that a preview of the OpenAI large language model known as o1 outperformed human physicians in multiple tests of clinical and diagnostic reasoning. Using case studies, o1 included the right diagnosis 78 percent of the time, while physicians were successful about 30 percent of the time. In another test, o1’s average score was 89 percent, compared to 34 percent for human doctors. The study, published on April 30 in Science, also found that o1 outperformed two expert attending physicians in diagnosing emergency room cases, identifying the exact or very close diagnosis in 67.1% of cases, while the human doctors did so in 55.3 percent and 50 percent of cases. Other research shows AI-assisted mammograms could be better at detecting breast cancer, per a Swedish study published by The Lancet in January. An AI model developed by the Mayo Clinic detected pancreatic cancer an average of 475 days and up to three years earlier than clinicians did. The article notes that robo-docs are not likely to take over healthcare anytime soon, and that "humans should be the ultimate baseline," but suggests AI models could competently assist in a variety of tasks.