<![CDATA[Artificial Intelligence (AI) plays a significant role in education in the modern era. A platform called Classpoint AI (CAI) has been widely utilized across various educational levels and disciplines worldwide. Numerous studies have reported positive responses toward using CAI as an interactive learning medium and an automatic question generator based on Bloom’s Taxonomy. This study aims to analyze the quality of test items generated by CAI quantitatively. The accuracy of a test item must be empirically validated to ensure high-quality learning. A non-probability sampling technique, specifically total sampling, was employed to select 39 female students in Phase F as research participants. Data were collected through a test consisting of 14 multiple-choice questions created by CAI to assess reading comprehension proficiency in Arabic. An Ex-post facto method was used in this study, as the researcher did not modify the pre-programmed CAI system. Three software tools (SPSS, Anbuso, and Anates) were used to analyze the item quality, and teachervalidation was also included. The findings revealed that most of the test items generated by CAI were invalid and exhibited low reliability. The questions were predominantly classified as easy. CAI-generated items demonstrated poor discrimination power butcontained effective distractors. For foreign language learning, CAI-generated questions need to reconsider the complexity of the vocabulary used. Considering the widespread use of this platform among educators on an international scale, this study contributes a global perspective by highlighting the need for both developer evaluation and educator discernment in utilizing CAI.]]>
Copyrights © 2025
