Qualitative thematic synthesis uncovers cognitive trade-offs and an autonomy paradox in personalized higher education, suggesting learning depth relies heavily on instructional design.
Key Points
To examine how generative and agentic artificial intelligence alter cognitive and motivational dynamics within personalized higher-education learning pathways.
Conducted a qualitative thematic synthesis of 33 purposively sampled and systematically coded studies using the Thomas and Harden framework.
Assessed evidence quality via a three-layer assurance framework measuring paper richness, synthesis-level rigor (composite score 0.890), and GRADE-CERQual confidence levels.
Identified three core themes: an Agency Shift toward proactive AI partnerships, a Cognitive Trade-Off pairing immediate performance gains with metacognitive risks, and an Autonomy Paradox where personalization both aids and constrains learner self-direction.
Developed a four-level typology of AI roles (reactive assistant, adaptive feedback system, semi-agentic planner, and fully agentic orchestrator), finding that educational depth is governed primarily by prompt literacy, assessment design, and academic integrity policies.