Algorithmic Mediation, Learner Agency, and Cognitive Autonomy in AI-Enhanced Online Learning Environments

Abstract

This article examines the psychological and behavioural implications of artificial intelligence integration in online learning environments, focusing on how algorithmic personalization, predictive feedback, and adaptive sequencing influence learner agency, cognitive autonomy, and self-regulation. Drawing on self-determination theory, cognitive psychology, and behavioural science, the study highlights both the supportive and constraining effects of AI-mediated learning. While adaptive systems can enhance perceived competence, reduce cognitive load, and support task alignment, they may also introduce subtle behavioural steering mechanisms that externalize regulation and diminish intrinsic motivation. The analysis shows that algorithmic nudging can recalibrate learners’ perceptions of control and responsibility, normalizing compliance with system-defined pathways. To address these tensions, the article proposes a psychologically grounded framework for ethically aligned AI-enhanced learning that balances adaptive scaffolding with reflective choice and exploratory engagement. By clarifying the cognitive and motivational mechanisms through which algorithmic systems reshape learner agency, the study offers design principles for autonomy-supportive AI-mediated learning environments.