Higher education is increasingly integrating artificial intelligence (AI) and sustainability-oriented learning, aligning with the United Nations Sustainable Development Goals (SDGs). Nonetheless, there is limited empirical evidence explaining how cognitive processes, AI-supported learning (AI-SL), and SDG awareness (SDGA) jointly relate to learner engagement (ENG) and learning outcomes (LO) across academic disciplines. This study examines the associations among working memory (WM), metacognition (MET), reasoning ability (REA), AI-SL, SDGA, ENG, and LO in higher education. Survey data were collected from undergraduate students across nine campuses of a public university system in Thailand. Multi-group structural equation modeling (MG-SEM) was employed to compare students enrolled in STEM and social science programs. Measurement model evaluation indicated satisfactory reliability and convergent and discriminant validity, with partial scalar invariance supported across groups, enabling meaningful structural comparisons. The results showed that ENG was strongly associated with LO in both groups. MET emerged as the strongest cognitive correlate of ENG overall, while REA was more strongly associated with ENG among STEM students, and SDGA showed a stronger association among social science students. AI-SL was positively associated with ENG and LO, with discipline-specific differences in the strength of the effect. The model explained substantial variance in ENG (R2 = 0.48) and LO (R2 = 0.52). These findings emphasize the value of integrating cognitive processes (WM, MET, and REA), AI-SL, and sustainability-oriented content within discipline-sensitive instructional designs in higher education.
Putthidech et al. (Thu,) studied this question.