The Impact of Generative AI on Critical Thinking: Self-Reported Cognitive Effort and Confidence Effects in High School Students

Students in secondary schools utilize generative artificial intelligence quite frequently. This paper aims to investigate the effects on students' learning processes, critical thinking, and cognitive effort. The study examines how high school students use AI tools for academic assignments and the connections between effort investment, critical thinking, trust, perceived learning, self-confidence, and confidence in AI. A poll was administered to twenty students in grades 9-12, who also reported using AI frequently. Likert-scale measurements of confidence, effort, trust, and learning perspectives, as well as assessments in the areas of critical thinking and qualitative remarks, were collected. Spearman's rank-order correlations and descriptive statistics showed that for our set of self reported data, higher levels of self-confidence are significantly correlated with better critical thinking and greater effort to comprehend and improve AI outputs. It also hints at higher estimates of self evaluation of generated content during AI use. Increased confidence in AI was associated with less evaluative effort, less checking behavior, and greater trust. The results, although being constrained by a smaller number of participants, pointed at a relationship of cognitive offloading and themes of AI serving as a cognitive scaffold for self-assured users. These results imply that the redistribution of cognitive effort brought on by generative AI is moderated by self-confidence. This paper therefore, highlights the importance of promoting self-control and metacognitive awareness.