Developing an AI-integrated mathematics learning model to enhance digital literacy: evidence from bilingual and non-bilingual contexts

Abstract

This study addresses the pedagogical gap between artificial intelligence (AI) integration and digital literacy development by developing and validating the artificial intelligence–integrated mathematics learning model (PMAI-e). Employing the Plomp research and development (R&D) framework, the study produced a comprehensive instructional model validated by experts with a high validity score (Avg. 3.64) and confirmed as practical by users. To evaluate effectiveness, a quasi-experimental study was conducted with 163 undergraduate mathematics students selected via cluster random sampling. Participants were divided into three groups: experimental group 1 (bilingual), experimental group 2 (non-bilingual), and a control group using conventional project-based learning. Data were analyzed using N-Gain, analysis of variance (ANOVA), and Tukey honestly significant difference (HSD) post-hoc tests. The results demonstrated that the PMAI-e model significantly enhanced digital literacy skills compared to the control group (p<0.001). The bilingual group achieved the highest improvement (N-Gain=0.71), followed by the non-bilingual group (N-Gain=0.67), with both significantly outperforming the control group (N-Gain=0.34). Notably, the two experimental groups differed significantly, according to post-hoc analysis (p=0.012), suggesting that integrating bilingual instruction with AI-supported pedagogy offers additive benefits to digital literacy development. The PMAI-e model is recommended as a scalable framework for fostering ethical and technical AI competencies in higher education.