ChatGPT-4 performance in rhinology: A clinical case series

Keypoints: Chatbot Generative Pre-trained Transformer (ChatGPT)-4 indicated more than twice additional examinations than practitioners in the management of clinical cases in rhinology. The consistency between ChatGPT-4 and practitioner in the indication of additional examinations may significantly vary from one examination to another. The ChatGPT-4 proposed a plausible and correct primary diagnosis in 62.5% cases, while pertinent and necessary additional examinations and therapeutic regimen were indicated in 7.5%–30.0% and 7.5%–32.5% of cases, respectively. The stability of ChatGPT-4 responses is moderate-to-high. The performance of ChatGPT-4 was not influenced by the human-reported level of difficulty of clinical cases.