As the popularity and the usage of the artificial intelligence (AI) tools is increasing in medical education, it is important to critically evaluate these resources and confirm their reliability. The current study proposes to assess the reliability and effectiveness of ChatGPT 3.5 and 4 for gross anatomical information on scalenovertebral triangle. ChatGPT versions 3.5 and 4 AI tools were used to explore the anatomical information on scalenovertebral triangle eight times on different days. The responses were qualitatively compared to the actual anatomy of the region and comments were made by the authors for each response. The replies given by ChatGPT were not appropriate (either incorrect, partially correct or incomplete) in any of the conversations. There was no major difference between the accuracy of responses, while comparing ChatGPT 3.5 and 4. Almost three out of four times, ChatGPT confused scalenovertebral triangle with scalene or interscalene triangle. None of the responses provided by ChatGPT 3.5 and 4 across all eight instances aligned even once with the standard anatomical description of the scalenovertebral triangle. A novice medical student may not be able to judge the difference between correct and incorrect, consequently may wrongly interpret the anatomy. So cautious planning and educator check is important while it is used. Further development and modifications of this AI tool are required to increase its potential to be used in medical education and healthcare.
Read full abstract