2024-01-04 10:37:35.AIbase.4.7k
ChatGPT-4 Has an Error Rate of Up to 83% in Pediatric Disease Diagnosis
Recent research shows that ChatGPT-4 has an accuracy rate of only 17% in pediatric medical cases, which is lower than the 39% accuracy in general medical cases. Researchers point out that ChatGPT struggles to identify disease relationships and suggest specialized training and provision of accurate medical literature to improve the model's accuracy. The model correctly diagnosed only 17 out of 100 pediatric cases, with most errors concentrated in the same organ system.
