AI-supported classroom planning, student engagement, and EFL achievement

This study examines the association between AI-based classroom planning and classroom climate and English as a foreign language (EFL) achievement among 240 secondary school students in West Bengal, India, with student engagement as a mediating variable. A single-group pretest–posttest design was employed across a three-week AI-supported intervention. Using SPSS 26 and AMOS 26, paired-sample t tests revealed statistically significant pre to-post improvements in classroom climate (t(239) = −76.90, p < .001, d = −4.96), student engagement (t(239) = −64.46, p < .001, d = −4.16), and EFL achievement (t(239) = −61.77, p < .001, d = −3.99). These effect sizes are unusually large and should be interpreted cautiously given the single-group design, short intervention duration, and potential testing and novelty effects. Structural equation modeling (SEM) confirmed partial mediation via student engagement (β = .479, p = .001, 95% CI [.320, .641]), explaining 63.3% of the variance in EFL achievement. These findings offer a scalable, evidence-based model for AI-supported language pedagogy in underresourced, multilingual EFL contexts in the Global South.