Educational Assistant: A Systematic Review of AI Tools, Protective Learning Strategies and Psychological Processes

Abstract

This systematic literature review examined how AI tools can be used effectively in educational settings by identifying psychological processes and learning strategies that facilitate positive AI-enhanced learning while preventing cognitive dependency and over-reliance. Following PRISMA guidelines, a comprehensive search was conducted across peer-reviewed journals from Springer, MDPI, and Elsevier published between 2020-2025. Studies were selected using Boolean search terms combining "AI" OR "artificial intelligence" with "learning strategies," "chatbot," "cognitive development," and related terms. Additionally, qualitative thematic analysis was employed to synthesize findings across three research questions addressing AI's effects on cognitive development, protective psychological processes, and design frameworks for safe AI learning environments. The analysis revealed three major themes regarding AI's effects: cognitive dependency characterized by cognitive offloading, individual variability in cognitive responses moderated by psychological needs satisfaction and metacognitive skills, and multidimensional impacts across cognitive, emotional, and motivational dimensions. Furthermore, four key psychological processes were identified that facilitate positive AI use: self-regulated learning strategies maintaining cognitive autonomy, metacognitive awareness enabling conscious monitoring of AI use, cognitive flexibility preventing dependence patterns, and intrinsic motivation through satisfaction of basic psychological needs (autonomy, competency, relatedness). Lastly, three design principles emerged for transforming AI tools: pedagogically-informed systems grounded in constructivist and activity theory, transparent and explainable AI systems improving metacognitive accuracy, and human-centered collaborative designs. As such, the paper shows that AI's impact on education depends on learner characteristics, AI design, and psychological processes rather than the technology itself. The review provides a framework for harnessing AI's potential without risking cognitive dependency, emphasizing that effective AI integration ORIGINAL ARTICLE requires strategic learning approaches, metacognitive awareness, and theoretically-grounded design principles. The findings suggest that AI's effects rely on contextual factors and implementation strategies rather than being inherently beneficial or detrimental.