İnfluenza ile İlgili Sıkça Sorulan Sorulara Yapay Zekâ Tarafından Üretilen Yanıtların Doğruluğu ve Hasta Yönlendirmesi

Objective: The aim of this study was to evaluate the responses generated by an artificial intelligence language model (ChatGPT-5.2) to the World Health Organization's influenza-related frequently asked questions in terms of relevance, accuracy, clarity, and completeness. Material and Methods:This was a cross-sectional observational study in which influenza-related FAQs obtained from the World Health Organization were submitted to ChatGPT-5.2.All questions were entered in a zero-state environment through individual sessions.The responses generated by ChatGPT-5.2were independently evaluated by 20 physicians with expertise in influenza and infectious diseases using a 10-point scale based on four criteria: relevance, accuracy, clarity, and completeness.Statistical analyses were performed, and inter-rater reliability was assessed using the intraclass correlation coefficient (ICC).Results: Overall, the assessment scores were highly favorable.The mean scores for relevance, accuracy, clarity, and completeness were 9.50±0.48,9.46±0.51,9.60±0.43,and 9.21±0.66,respectively.Inter-rater reliability analysis demonstrated a moderate and statistically significant agreement for completeness (ICC=0.581,p