Impact of Generative AI Integration on Design Sketching Pedagogy in Higher Education: An Empirical Study in Mianyang, China

Abstract

This study aims to address teaching bottlenecks in design sketching courses, such as insufficient creative divergence and high technical barriers. It introduces an AI agent (Doubao) as an auxiliary tool in university teaching and employs a mixed-methods approach combining quantitative questionnaires (N=486), student interviews (N=35), and expert evaluations (N=3) to systematically assess its application effectiveness. Results indicate that AI effectively enhances students' creative generation efficiency, visual expression capabilities, and learning confidence, demonstrating particularly significant effects in conceptual divergence and creative process restructuring. However, it also brings issues such as insufficient operational understanding, stylistic homogenization, and creative dependency. Expert evaluations affirm its value in stimulating innovation and expanding visual language while cautioning against the risks of over-reliance. The study confirms that AI agents possess dual-natured pedagogical potential, requiring an application model centered on “human-machine collaboration with human-led control” for effective integration. This research provides empirical evidence and practical guidance for AI applications in art and design education, as well as for sketching pedagogy reform.