Reply to Comment on “Evaluation and Comparison of the Knowledge Levels of Current Artificial Intelligence Programs on Retinal/Vitreous Diseases and Treatment Methods”

Abstract

Dear Editor, We would like to thank Dr. Shrivastava for his positive and developing criticisms regarding our article and their close interest in us. We tested 46 questions questioning retina and vitreous diseases and treatment methods in our study.1 These questions consisted of questions in the book of BCSC, a respected book series of the American Academy and Association of Ophthalmology, Retina, and Vitreous. This book series is an important ophthalmology source with a certain level of international reputation and up-to-date information. We included all the study questions in this book in our study and ensured the homogeneity of the questions in this way. In addition, since adding additional questions could disrupt the homogeneity of the questions, we did not add questions from other sources in our study. However, we believe that studies to be conducted with larger and more up-to-date series may be effective in eliminating various contradictions on this subject. We previously provided the questions to be answered only after the artificial intelligence programs gave the command to indicate the correct answer option. We also eliminated the interpretations of the answers by comparing the given answer options with the answer key at the back of the book and accepted their success in providing a more accurate and clear answer as a criterion. In this way, we eliminated the interpretations in accepting the questions as right or wrong. We cleaned up the chat sessions after each question, which we adopted to break the memory effect and to keep this context-based direction changes in answering the questions to a minimum.1 The Pearson Chi-square test is a test used to evaluate categorical data. In 3 × 2 groups, when the expected value is below 5 and the porosity rate is below 20%, the Pearson Chi-square test is used. In 2 × 2 porosities, when the expected minimum value is between 5 and 25, the Yates Chi-square test is used. These criteria were applied by providing the conditions in our study.1 The use and dissemination of artificial intelligence programs in retina and vitreous diseases still have many limitations. Testing these limitations in a larger series of studies is of great importance in identifying the shortcomings of artificial intelligence programs and in the development of these programs. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.

Ähnliche Arbeiten