AI-Based Learning Environments and Their Impact on Academic Success: The Mediating Role of Engagement and Instructional Design

Abstract

This study investigates AI-based learning environments and their influence on the academic success of Grade 12 Biology students at Islamabad Model College for Boys, F-10/3 (IMCB F-10/3), Islamabad, Pakistan, with specific attention to the mediating roles of student engagement and instructional design quality. Utilizing a quasi-experimental pretest-posttest control group design, a stratified random sample of N = 80 Grade 12 Biology students was drawn from the institution's four sections and assigned to experimental (n = 40) and control (n = 40) conditions. The experimental group received Biology instruction delivered within AI-enhanced learning environments—incorporating adaptive learning platforms, intelligent feedback systems, and AI-powered content visualization—structured through principled instructional design frameworks (ADDIE and UDL). Standardized data collection instruments included the Academic Success Assessment (ASA), the Multidimensional Student Engagement Scale (MSES), and the Instructional Design Quality Rating (IDQR). Statistical analyses comprised independent samples t-tests, paired samples t-tests, one-way ANOVA, and Pearson correlation with mediation analysis. Findings demonstrated statistically significant academic success gains in the experimental group (t(78) = 5.12, p < .001, d = 1.14) and confirmed that student engagement (β = .43, p < .001) and instructional design quality (β = .37, p < .001) jointly and significantly mediated the relationship between AI-based learning environments and academic success. The study concludes that AI-based environments, when structured through deliberate instructional design and leveraged to sustain multidimensional engagement, produce substantially superior academic outcomes in secondary Biology education.