Harvard Study Finds OpenAI's o1 AI Matches or Beats ER Doctors in Triage and Diagnosis

OpenAI's o1 AI model matches or outperforms ER doctors in triage, diagnosis, and clinical case management across 76 real emergency cases, according to a Harvard-led study published in Science, with researchers urging urgent clinical trials to explore how AI can partner with physicians to reduce errors and improve patient outcomes.

• A Harvard-led study published in Science reveals that OpenAI's o1 preview AI model matches or outperforms human physicians in emergency room triage, diagnosis, and clinical case management across 76 real ER cases at a Boston hospital.
• The AI proves especially strong in diagnosing rare and complex diseases, excelling on benchmark cases from Massachusetts General Hospital, and significantly outpaces both previous AI models and humans aided by Google search on management reasoning tasks.
• Researchers stress that AI is not replacing doctors, but call for urgent clinical trials to determine how the technology can best partner with physicians to reduce diagnostic errors, provide second opinions, and ultimately improve patient outcomes.