A new study from Harvard Medical School reveals that AI models outperformed human doctors in diagnosing patients in emergency rooms. The o1 model offered the exact diagnosis in 67% of cases, compared to only 50-55% for human physicians.

Researchers at Harvard and Beth Israel Deaconess Medical Center found that AI models were particularly effective during initial triage, where time is critical. However, they caution that more real-world testing is needed before fully integrating AI into clinical practice.

The study did not claim that AI could replace human doctors in making life-or-death decisions but highlighted the urgent need for further evaluation and regulation of these technologies. Critics argue that current models are limited to text-based information and may struggle with non-text inputs like images or lab results.

Doctors warn that there’s no formal framework for accountability around AI diagnoses, and patients still prefer human guidance in critical situations. Emergency physicians emphasize that their primary goal is not just diagnosing but ensuring patient safety through timely interventions.