Examining AI-Enhanced Bilingual Scaffolding in Educational Technology English: Evidence from a Quasi-Experimental Study

Abstract

This study examined AI-enhanced bilingual scaffolding in Educational Technology English within an ESP framework. It investigated whether strategic bilingual mediation and AI-assisted feedback were associated with higher student engagement and higher-order learning than traditional instruction. Using a quasi-experimental pretest-posttest non-equivalent control group design, the study involved 84 undergraduate students from an Indonesian Educational Technology program. Two intact classes were assigned as experimental and control groups. Data were collected through a student engagement scale, higher-order learning performance assessment, classroom observation checklists, and reflection logs. The data were analyzed using descriptive statistics, independent-samples t-tests, gain-score analysis, effect-size estimates, and thematic interpretation. The findings showed that the experimental group achieved greater gains in cognitive engagement (gain = 1.27; d = 1.46) and creation skills (gain = 18.7; d = 1.52), with significant between-group differences (p < .001). Qualitative findings suggested that AI functioned as a cognitive partner, supported bilingual conceptual transfer, reduced anxiety, and increased linguistic confidence. The study contributes to scaffolding and bilingual pedagogy by proposing an AI-enhanced bilingual co-scaffolding model for ESP curriculum design, AI literacy integration, and lecturer professional development.