<![CDATA[This study examines the cognitive and pedagogical mechanisms of AI-integrated bilingual English for Specific Purposes (ESP) instruction, transcending traditional technology adoption viewpoints to elucidate how instructional aids influence learning outcomes. The research examined the impact of AI Pedagogical Support and Bilingual Scaffolding on ESP Learning Performance, mediated by Student Engagement and Academic Self-Efficacy. A quantitative explanatory cross-sectional approach was utilized, involving 312 undergraduate ESP students participating in an AI-assisted bilingual English for Educational Technology course at an Indonesian institution. Data were gathered using a validated 5-point Likert-scale questionnaire assessing five latent constructs and analyzed via PLS-SEM with SmartPLS 4, encompassing reliability, validity, structural, mediation, and predictive model evaluations. The results indicated that AI Pedagogical Support was a significant predictor of Student Engagement (β = .41) and Academic Self-Efficacy (β = .36), whereas Bilingual Scaffolding significantly predicted Student Engagement (β = .29) and Academic Self-Efficacy (β = .31). Both mediators substantially impacted ESP Learning Performance, resulting in R² = .62, with complementing partial mediation validated by strong indirect effects. The findings indicate that AI enhances ESP outcomes mainly by influencing learner cognition, confidence, and pedagogically facilitated bilingual support, thereby providing a significant theoretical contribution to a cognitive–pedagogical ecosystem model and practical recommendations for the redesign of AI-integrated ESP curricula in higher education.]]>
Copyrights © 2026
