Predicting AI Readiness and Teacher Motivation in Blended Secondary Mathematics Classrooms

Abstract This study investigated the structural relationships between technology-supported instructional tools and teachers' perceptions and motivation among 53 purposively sampled secondary school teachers in Central Ethiopia. While raw data deviated from normality (p < 0.001), stable homoscedasticity and safe collinearity diagnostics (VIF range: 1.147–1.249; maximum Condition Index: 20.689) justified ordinary least squares multiple linear regression. The overall regression model was highly significant (F(3, 49) = 22.648, p < 0.001), explaining 58.1% of the total variance in teacher perception and motivation (R^2 = 0.581, Adjusted R^2 = 0.555). Individually, the Use of Adaptive Learning Platforms was the strongest unique positive predictor (β = 0.683, t = 6.611, p < 0.001), followed by AI-Enhanced Feedback Systems (β = 0.261, t = 2.641, p = 0.011). Conversely, Interactive Assessment Tools yielded a non-significant effect (β = -0.146, p = 0.155). Descriptive analyses revealed a pedagogical paradox. Teachers showed strong baseline literacy regarding blended learning concepts (M = 3.72) and GeoGebra (M = 3.85), but operational usage of Google Workspace sat below the neutral midpoint (aggregated M = 2.896). Usage clustered in passive workflows like Gmail (M = 3.19) and YouTube (M = 3.08), while active content-creation systems like Google Forms (M = 2.77) and Slides (M = 2.74) lacked high-frequency deployment. Systemic policy under the Digital Ethiopia 2025 initiative must pivot professional development toward active digital content curation and algorithmic orchestration.