Artificial intelligence in anatomy education: a systematic review of ChatGPT’s effectiveness as a learning tool

Abstract

Background: This systematic review evaluated the reliability of ChatGPT as a supportive tool in anatomy education and its performance across various anatomical topics. Following PRISMA guidelines, English-language studies published between 2023 and 2025 were included. Searches were conducted in Google Scholar, PubMed, Springer, and Scopus using the keywords "ChatGPT," "anatomy," and "education." Results: After removing duplicates and screening titles and abstracts, 26 full-text articles were reviewed. Quality assessment using Joanna Briggs Institute criteria rated 22 articles as moderate quality, 1 as high quality, and 3 as low quality; low-quality studies were excluded, leaving 23 articles for analysis. Accuracy rates ranged from 75.2% to 98.6% for ChatGPT-4o, 4.0% to 97.3% for ChatGPT-4, 28.0% to 72.0% for ChatGPT-3.5, and 34.2% to 67.2% for unspecified models. ChatGPT-3.5 Turbo showed 38.4% accuracy in one study, and ChatGPT-4o Mini, 89.2%. Gross anatomy was the most addressed topic (29.0%), followed by thoracic (9.7%), abdominal, upper extremity, and musculoskeletal anatomy (6.5%). ChatGPT-3.5 was the most used model (46.4%), followed by ChatGPT-4 (29.6%), ChatGPT-4o (7.1%), and ChatGPT-4o Mini (7.1%). Conclusion: ChatGPT has potential as a complementary tool in anatomy education, yet careful evaluation and further evidence-based research are needed to improve its reliability and effectiveness.

Ähnliche Arbeiten