Translated data: The latest research indicates that ChatGPT-4 achieves an accuracy rate of only 17% in pediatric medical cases, performing worse than in general medical cases last year. Researchers suggest that training and providing accurate medical literature could potentially enhance ChatGPT's accuracy in pediatric diagnosis. The study emphasizes the irreplaceable role of human pediatricians' clinical experience in the medical field. Out of 100 cases, the model correctly diagnosed only 17, with most errors concentrated within the same organ system. Training with specific medical data is expected to improve the accuracy of large language model-based chatbots in diagnosing pediatric diseases.