AI Sets a New Benchmark in ER Triage, But Doctors Aren’t Replaced
TL;DR
Harvard/BIDMC researchers tested OpenAI’s o1-preview against two physicians in ER triage; the AI diagnosed correctly in 67.1% of 76 real cases (physicians 55.3% and 50%), and offered a helpful differential in 97.9% of cases. The study suggests AI could complement clinicians, not replace them, but notes limitations with multimodal data and hallucinations, underscoring the need for rigorous trials and safety safeguards before broad adoption.
- AI Just Beat Doctors at Diagnosing ER Patients. Don’t Get All Excited Gizmodo
- In real-world test, an AI model did better than ER doctors at diagnosing patients NPR
- AI outperforms doctors in Harvard trial of emergency triage diagnoses The Guardian
- Paging Dr. AI to the ER? Artificial intelligence shows promise in emergency room diagnosis CBC
- As artificial intelligence shows off diagnostic chops, scientists reckon with the way forward statnews.com
Want the full story? Read the original reporting
Read on Gizmodo