Beyond the Instrumental Use of AI to foster Mathematics students’ Metacognitive Awareness and Critical Thinking

The contemporary challenge of integrating Artificial Intelligence (AI) in education [1--5] highlights the need for a pedagogical approach that goes beyond mere instrumental use, focusing instead on fostering metacognitive and critical thinking skills [6]. Ennis's framework [7] and Flavell's theory [8] explain how learners develop self-regulation and critical evaluation of their own problem-solving strategies. In mathematics education, Schoenfeld's seminal work [9] further emphasizes the importance of metacognitive awareness in mathematical problem solving, highlighting the necessity for students to monitor and regulate their mathematical thinking processes. More recently, in mathematics education research, Contel and Cusi [10] and Miranda [11] have explored how ChatGPT supports metacognitive processes and contributes to shaping mathematical identity during university-level problem solving. The presence of AI as a collaborative "participant," sharing its solution after students complete their work, stimulates critical thinking, and aligns with Vygotsky's [12] view that interaction with external tools fosters knowledge construction and internalization of higher cognitive processes. This pilot study, as part of a broader interdisciplinary project investigating the metacognitive potential of AI in education, examines how structured AI integration in mathematical problem solving can foster: (1) metacognitive competencies in university mathematics students; (2) critical evaluation capabilities of problem-solving strategies; and (3) deeper conceptual understanding through comparison of human- and AI-generated solutions.