Studies published in peer-reviewed medical journals

GPT-4 AI Surpasses Doctors in Diagnostics Across Specialties, Equals Radiologists

Apr 17, 2024, 11:29 PM

Recent studies have highlighted the capabilities of AI in the medical field, particularly in diagnosing and assessing various health issues. GPT-4, an advanced AI model developed by OpenAI, has demonstrated proficiency in several medical specialties, surpassing or matching the diagnostic accuracy of human doctors in some cases. Notably, GPT-4 has excelled in psychiatry, general surgery, internal medicine, and notably in ophthalmology, where it performed comparably to expert ophthalmologists. The AI's performance in pediatrics and OB/GYN was slightly lower but still above many practicing physicians. This technology has also passed licensing board exams in multiple medical specialties, showcasing its potential to support and enhance patient care through more accurate diagnostics and efficient treatment suggestions. Additionally, GPT-4V, which includes image analysis, has shown to equal radiologists in identifying radiology report errors and surpassed pathologists in diagnosing metastatic cancer cells.