Integrating Artificial Intelligence Into Mathematics Learning to Develop Students Critical Thinking and Metacognitive Skills

Abstract

The integration of artificial intelligence (AI) into mathematics learning offers opportunities to strengthen students’ critical thinking and metacognitive skills, but its educational value depends on how students interact with and evaluate AI-generated information. This study investigates the effects of AI-supported mathematics learning on students’ critical thinking and metacognitive skills and examines student–AI interaction patterns during mathematical problem solving. An explanatory sequential mixed-methods design was employed involving 60 secondary school students participating in structured AI-supported mathematics learning activities. Quantitative data were obtained from critical-thinking and metacognitive assessments administered before and after the intervention, while qualitative data were collected through classroom observations, student interviews, written solution processes, and 3,842 AI interaction turns. The findings indicate substantial improvements in critical thinking, with the overall mean score increasing from 52.49 (SD = 8.73) to 73.37 (SD = 8.21), accompanied by a significant difference, t(59) = 39.36, p < .001, and a very large effect size (Cohen’s d = 2.46). Metacognitive performance also increased from 2.90 (SD = 0.43) to 3.85 (SD = 0.40), t(59) = 40.34, p < .001, with Cohen’s d = 2.60. Explanation-seeking, strategy exploration, verification, critical questioning, and reflection were prominent interaction patterns. Verification was positively associated with learning gains, whereas direct answer-seeking showed negative associations. These findings indicate that AI can function as a cognitive and metacognitive scaffold when students retain responsibility for verification, reflection, and mathematical judgment.