Effectiveness of Artificial Intelligence in Higher Education: A SEMANN Approach from Students’ Perspectives

Abstract

This study explores the effectiveness of Artificial Intelligence (AI) in higher education, focusing on Accuracy and Reliability (AR) and Responsiveness (RS) as determinants of Overall Effectiveness of AI (OEAI). Employing a dual analytical approach of Structural Equation Modeling (SEM) and Artificial Neural Network (ANN) analysis, for the validate of two hypotheses. H1: establishes a significant positive impact of AR on OEAI, supported by a path coefficient (β = 0.706; t = 15.097). H2: confirms a positive influence of RS on OEAI, albeit with lower values (β = 0.119; t = 2.466), suggesting a lesser impact on students’ overall perception. Reliability and validity analyses ensure the robustness of this research model, validating convergent and discriminant validity. Neural network analysis underscores the importance of AR and RS, with AR identified as the most influential factor (100%), followed by RS (68%). The model exhibits high accuracy (R² = 0.895), confirming its predictive power.