Abstract
Cultivating critical thinking skills remains a foundational objective in undergraduate philosophy education, yet providing frequent, personalized, and dialectical feedback at scale poses severe pedagogical challenges. This quasi-experimental study evaluated the efficacy of a customized Generative Artificial Intelligence (GenAI) formative feedback system designed to stimulate Socratic inquiry and argumentative rigor. Across a 14-week semester, undergraduate students enrolled in four sections of an introductory philosophy course (N = 142) were assigned to either an experimental group receiving scaffolded GenAI feedback on argumentative draft essays (n = 72) or a control group receiving traditional instructor-mediated rubric feedback (n = 70). Critical thinking was assessed using the California Critical Thinking Skills Test (CCTST) alongside rubric-based evaluations of philosophical writing, focusing on premise acceptability, deductive validity, and counterargument integration. Analysis of covariance (ANCOVA) revealed that students utilizing the GenAI feedback system achieved statistically significant gains in overall critical thinking (F(1, 139) = 14.82, p < .001, partial η² = .096), with marked improvements in deduction and evaluation subscales. Textual analysis of essay revisions indicated that the experimental cohort engaged in deeper dialectical revisions, producing more robust anticipations of opposing viewpoints. These findings suggest that deliberately scaffolded, dialogue-oriented GenAI feedback can effectively augment cognitive engagement and argumentative sophistication in the humanities.
