VIRTUAL MADRID FOCUS ON PEDAGOGY

Abstract

The arrival of generative artificial intelligence (GenAI) in design education has intensified long-standing debates about authorship, critique, and the nature of creative judgement. This study investigates how design students interpret and evaluate AI-generated feedback compared with traditional teacher critique, focusing on perceived usefulness, trustworthiness, and creative value. Using a mixed-methods design, 504 design students in Bangalore completed a 17-item survey combining Likert-scale measures with open-response reflections. Quantitative analysis employed descriptive statistics, paired t-tests, correlations, ANOVA, and effect sizes, while computational validation used histogram correlation and structural-similarity metrics (SSIM) to examine distributional alignment between AI and teacher ratings. Results show that AI feedback is valued for speed, clarity, and ideation support yet rated substantially lower in trust (t(461) = –14.22, p < .001, dz = 0.88). Helpfulness strongly predicted intention to reuse AI feedback (r = .68). Thematic analysis of approximately 6,000 coded meaning units identified three clusters of concern—Efficiency (Immediacy, Ideation, Accuracy), Relationality (Empathy, Objectivity, Hybrid Preference), and Existential Impact (Ethics, Agency, Anxiety)—mapping a complex and evolving feedback ecology in design studios. SSIM values (.91 for helpfulness; .63 for trust) confirmed that cognitive convergence in perceived usefulness coexists with affective divergence in trust. Together, the findings conceptualize AI feedback for design students as cognitively efficient but relationally incomplete, and propose a hybrid “co-creative feedback ecology” in which AI augments, rather than replaces, the interpretive and empathetic functions of human critique.