Work

Thinking About Thinking: Metacognition, AI, and the Future of Self-Evaluation in Learning

Year: 2026
Type: article
Study design: Systematic Review
Abstract: Metacognition, the ability to "think about thinking" is a fundamental psychological construct that predicts academic achievement and lifelong learning. In the artificial intelligence (AI) era of 2026, metacognition faces transformative evolution as AI-powered mobile applications emerge to facilitate personalized self-evaluation and self-regulation. However, integrating AI into metacognitive processes raises critical questions regarding epistemic sovereignty humans' ability to maintain agency over their own cognitive processes. This paper aims to synthesize findings from 10 recent studies (2025-2026) on metacognition in the AI era, with particular focus on mobile app-based self-evaluation mechanisms and their pedagogical and psychological implications. Methods: A systematic literature review and conceptual analysis was conducted on 10 peer-reviewed studies published between 2025-2026. Search strategy employed Open Alex API with keywords "metacognition" AND ("artificial intelligence" OR "AI") AND ("self-evaluation" OR "self-regulation"). Thematic analysis identified convergent findings, divergent debates, and research gaps. Synthesis identified four convergent themes: (1) metacognition as a multi-dimensional construct with equally important monitoring and controlling components; (2) AI and mobile technologies as metacognitive scaffolds that enhance awareness through real-time feedback and personalization; (3) self-evaluation accuracy as a critical mediator for effective self-regulation; and (4) a lifespan development perspective showing that metacognitive skills develop throughout life and can be supported by AI tools. Analysis also identified three divergent debates: individual vs. shared metacognition, AI as tool vs. agent, and metacognitive measurement approaches. AI integration in metacognitive processes offers significant opportunities to enhance metacognitive awareness and self-regulation, yet requires careful consideration of epistemic sovereignty and metacognitive offloading risks. This paper recommends developing AI-mediated metacognitive interventions that preserve human agency and prioritize epistemic sovereignty principles. Practical implications for educators, EdTech developers, and policymakers are comprehensively discussed.
Language: English
Field: Psychology
Open Access status: hybrid
OpenAlex ID: W7225927051
FWCI:
Cites: 4
Cited by:
Related to: 10