Are ChatGPT’s Answers to Questions About Salivary Gland Diseases Accurate and Reliable?
Turkish Archives of Otorhinolaryngology, cilt.63, sa.3, ss.143-150, 2025 (ESCI, Scopus, TRDizin)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 63 Sayı: 3
- Basım Tarihi: 2025
- Doi Numarası: 10.4274/tao.2025.2025-1-17
- Dergi Adı: Turkish Archives of Otorhinolaryngology
- Derginin Tarandığı İndeksler: Emerging Sources Citation Index (ESCI), Scopus, TR DİZİN (ULAKBİM)
- Sayfa Sayıları: ss.143-150
- Anahtar Kelimeler: Salivary gland, salivary gland tumor, sialadenitis, generative artificial intelligence, medical education, professional practice
- Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
- Sağlık Bilimleri Üniversitesi Adresli: Evet
Özet
Objective: Chat Generative Pre-Trained Transformer (ChatGPT) is an artificial intelligence model that can generate human-like text dialogs to inputs. There are no ChatGPT studies in the literature only on salivary gland diseases. This study evaluates the accuracy and reliability of ChatGPT’s answers to questions in salivary gland diseases, focusing on its potential use in training otolaryngology professionals. Methods: Sixty-one questions, categorized as “basic knowledge,” and “salivary gland tumors” were posed twice using ChatGPT-4. Answers were categorized as 1 (completely correct and comprehensive), 2 (partially correct), 3 (misleading information containing correct and incorrect statements), or 4 (completely incorrect). The accuracy of the answers was evaluated by two ear, nose and throat specialists. Inconsistencies in the rating were resolved by a third reviewer. Reproducibility was assessed by the agreement between the first and second answers. Results: Fifteen (24.6%) of the questions were about basic knowledge, while 46 (75.4%) were about salivary gland tumors. ChatGPT gave “completely correct and comprehensive” answers to 54 (88.5%) questions and “partially correct” answers to seven (11.5%) questions. “Misleading information containing correct and incorrect statements” and “completely incorrect” answers were not received. The reproducibility rate of first and second answers was 97%. Conclusion: ChatGPT provided highly accurate and reproducible answers to questions about salivary gland diseases. ChatGPT is thought to be an important source of information for otolaryngology professionals. Although the results of our study show that ChatGPT is highly successful, more studies are needed in this field.