"Study Reveals ChatGPT's 17% Success Rate in Pediatric Medical Diagnoses"
A study published in JAMA Pediatrics found that ChatGPT had only a 17% accuracy rate when diagnosing pediatric medical cases, highlighting significant limitations in its current ability to replace human pediatricians. The AI struggled with understanding complex medical relationships and provided incorrect or overly broad diagnoses in the majority of the 100 cases tested. Researchers suggest that improvements could be made by training the AI on accurate medical literature and providing it with real-time medical data.