The Shock of 'Better-than-Doctor' Diagnostic Accuracy: Harvard Study Shows OpenAI’s o1 Outperforms Human Physicians in ER Triage
A groundbreaking study from Harvard University reveals that OpenAI’s o1 model achieved a 67% diagnostic accuracy rate in emergency room settings, significantly outperforming human triage doctors who averaged 50-55%. This shift marks a turning point in medical AI from 'assistant' to 'expert.'