An Analysis of the Effectiveness of Using the Generative Artificial Intelligence on Mathematical Word Problems: A Case Study of Fifth-Grade Remedial Instruction Students

Abstract

Mathematical word problems involve both mathematical operations and reading comprehension, which often pose a significant challenge for students with low academic achievement. This study integrates generative AI (Adaptive Learning Platform, ALP) into instruction, using its Socratic questioning approach to guide students in understanding problems and developing problem-solving skills in mathematics. The participants were fifth-grade remedial students from an elementary school. A one-group pretest-posttest design was adopted: during the first 9 weeks, instruction on mathematical word problems was delivered by the teacher using the Socratic questioning method; in the following 9 weeks, the generative AI (ALP) was incorporated into the learning process. The results showed that the use of generative AI (ALP) effectively enhanced students’ achievement in solving mathematical word problems. Overall, students' learning attitudes improved, with a notable reduction in learning anxiety, indicating that their anxiety when solving mathematical word problems significantly decreased.