Integrating AI Generative ChatBot Agent as Adaptive Intervention in Mathematics Learning for Elementary School Students

Mathematics learning in elementary school plays an important role in developing critical thinking through contextual problem solving. However, in practice, many students still experience difficulties in understanding information and determining appropriate solution strategies. Various problem-solving approaches have been developed, but the implementation of learning that integrates systematic thinking while accommodating individual differences in learning pace remains limited. Recent studies highlight the use of artificial intelligence (AI) to support student learning, as well as the development of adaptive learning to accommodate individual learning needs. However, the explicit integration of AI with adaptive learning is still underexplored. Therefore, this study aims to examine the use of a generative AI chatbot agent as an adaptive learning intervention to support students’ structured problem-solving processes.

This study employed a quantitative approach using a pre-experimental one-group pretest–posttest design. Participants were elementary school students selected through total sampling. The learning process followed structured problem-solving steps: Restate, Extract, Map, Solve, and Justify (REMSJ), to guide students’ thinking systematically. The results showed a significant improvement in students’ problem-solving ability (p < 0.001), with the mean score increasing from 125.82 to 173.88 and a large effect size (dz = 1.74). These findings indicate the effectiveness of the adaptive AI chatbot intervention in supporting structured mathematical problem solving. Future studies are recommended to examine AI-based adaptive learning in broader contexts and across different educational levels