ARTIFICIAL INTELLIGENCE AS A COGNITIVE PARTNER IN EDUCATION: THE TRANSFORMATION OF LEARNING AUTONOMY
ARTIFICIAL INTELLIGENCE AS A COGNITIVE PARTNER IN EDUCATION: THE TRANSFORMATION OF LEARNING AUTONOMY

Summary/Abstract: This paper examines artificial intelligence as a cognitive partner in education and analyzes how this partnership reshapes learning autonomy, defined as learners’ capacity to set goals, regulate effort, evaluate understanding, and assume responsibility for knowledge construction. The purpose is to clarify both the educational value and the unintended costs of delegating cognitive activities such as explanation, feedback seeking, drafting, and problem solving to artificial intelligence systems that interact with students in real time. The methodology is a structured narrative literature review that synthesizes recent peer-reviewed research and policy-oriented scholarship on artificial intelligence–mediated learning, with particular attention to autonomy, self-regulated learning, cognitive offloading, academic integrity, equity, and pedagogical design. The results indicate a dual pattern: artificial intelligence can strengthen autonomy when it is used as scaffolding that supports planning, monitoring, and reflection, yet it can weaken autonomy when it substitutes for productive struggle and reduces learners’ engagement in sense-making, verification, and revision. Evidence across the literature suggests that the direction of impact depends on task type, learner expertise, assessment conditions, and the transparency of the artificial intelligence system’s reasoning. The conclusions argue that autonomy is not automatically enhanced by personalization or immediate assistance; rather, autonomy is transformed into a distributed practice that requires explicit instructional framing and metacognitive guidance. Recommendations include designing learning activities that require students to justify decisions, compare artificial intelligence outputs with alternative sources, document learning processes, and reflect on when assistance supports understanding versus mere completion; at the institutional level, the paper recommends updating assessment practices, establishing clear ethical guidelines for classroom use, and supporting teacher professional development focused on pedagogical orchestration of artificial intelligence tools

• Details
• Contents