Work

Enhancing heritage-oriented design learning through a value-sensitive approach to generative AI in higher education aesthetic pedagogy

Year: 2026

Type: article

Abstract: The preservation of Intangible Cultural Heritage (ICH) often struggles with static archival methods and a growing disconnect from contemporary design practices. While generative artificial intelligence offers a rapid mechanism to translate traditional aesthetics into modern forms, off-the-shelf models frequently suffer from cultural bias and overrepresent Western-centric data, leading to inauthentic outputs. This study introduces a pedagogical framework combining Generative AI with Value-Sensitive Design (VSD) to address these challenges in higher education. Focusing on two regional heritages—Xiang Embroidery and Changsha Kiln ceramics—we conducted a 12-week quasi-experimental study with 118 design students. The experimental group utilized a custom AI pipeline employing multimodal LLM tagging, Low-Rank Adaptation (LoRA) fine-tuning, and ControlNet spatial constraints to ensure algorithmic outputs remained culturally accurate. Quantitative results revealed that students using the VSD-guided AI demonstrated statistically significant improvements in aesthetic appreciation, design implementation, and cultural identity compared to the control group using traditional methods. Qualitative feedback indicated that the AI pipeline effectively lowered technical drafting barriers, shifting students’ cognitive focus toward cultural curation and critical aesthetic judgment. The findings suggest that ethically constrained generative models effectively support heritage-oriented design learning in regional universities. Rather than attempting to “revitalize” the heritage itself, this structured approach provides students with a culturally grounded framework. It allows them to safely engage with traditional aesthetics and translate these elements into contemporary formats without risking cultural dilution.

Source: Scientific Reports

DOI: 10.1038/s41598-026-71408-2

Authors: Tingting Yu, Yixin Li

Institutions: Xiangtan Institute of Technology, Hunan University of Technology and Business

Language: English

Topic: Artificial Intelligence in Education

Subfield: Education

Field: Social Sciences

Domain: Social Sciences

SDGs: Quality education, Reduced inequalities

Open Access status: gold

Estimated APC paid: $2,690