A Multi-Path Direct Model of AI-Supported Learning: The Roles of Algorithmic Trust, Cognitive Load Management, and Engagement in Driving Learning Innovation Performance

Abstract

Aim/Purpose: This study addresses a limitation of existing AI-supported learning research: its reliance on linear models that fail to capture the complex interplay among technological, cognitive, and behavioral factors in shaping learning innovation performance. Background: To overcome this limitation, the study proposes a multi-path direct model that simultaneously examines the roles of AI usage, algorithmic trust, cognitive load management, engagement, and adaptability in a unified framework. Methodology: The study adopts a quantitative approach using Partial Least Squares Structural Equation Modeling. Data were collected from undergraduate engineering students with experience in AI-supported learning environments, and the analysis followed measurement and structural model evaluation procedures. Contribution: This study contributes to the literature by advancing a multi-path direct perspective that captures interdependent relationships among key constructs, redefines the roles of engagement and algorithmic trust, and positions self-regulated learning adaptability as the central mechanism driving innovation outcomes. Findings: The findings show that AI learning companion usage significantly strengthens algorithmic trust and learning adaptability, while cognitive load management enhances both engagement and adaptability. Self-regulated learning adaptability emerges as the strongest predictor of learning innovation performance. In contrast, engagement and algorithmic trust do not directly influence innovation outcomes, indicating that deeper cognitive transformation is required. Recommendations for Practitioners: Educators should emphasize adaptive learning strategies and critical use of AI tools while designing learning environments that support cognitive regulation rather than focusing solely on engagement. Recommendation for Researchers: Researchers should adopt integrative, non-linear models to capture the complexity of AI-supported learning better and to investigate indirect mechanisms among key variables further. Impact on Society: The study provides insights for developing more effective AI-enhanced learning systems that foster innovation, critical thinking, and adaptability in increasingly complex digital environments. Future Research: Future studies should extend this model across disciplines and cultural contexts, apply longitudinal designs, and incorporate objective performance measures to strengthen causal understanding.