Bridging Tradition and Innovation: Large Language Models for Cultural Heritage Garment Preservation and Sustainable Development

Abstract

Cultural heritage garments face unprecedented challenges in globalization, including loss of traditional craftsmanship, aging artisan communities, and difficulties in sustainable development. This study explores how Large Language Models (LLMs) can facilitate digital transformation while preserving cultural authenticity. We developed an LLM-driven framework encompassing four modules: craft knowledge extraction, culturally-informed design generation, artisan-centered supply chain coordination, and digital communication. The knowledge extraction module systematically preserves tacit craft knowledge by constructing a knowledge graph with 1,526 entities and 14,902 relations. The design module balances cultural preservation with contemporary aesthetics. The supply chain module supports artisan workshops through intelligent coordination systems. The communication module enhances cultural awareness through automated content generation. Empirical validation across five ethnic minority heritage garments demonstrates significant impacts: knowledge extraction achieved 92.3% precision, design quality reached 8.3/10, production efficiency improved 45.1%, artisan satisfaction increased 32.3%, and brand awareness grew from 28% to 67%. This research addresses critical sociocultural questions: How can technology empower rather than replace traditional artisans? How can we achieve sustainable development while maintaining cultural authenticity? Our findings suggest that appropriately designed AI systems can support intergenerational knowledge transfer, enhance artisan livelihoods, and foster cultural identity, providing practical pathways for achieving dual goals of cultural protection and economic sustainability in traditional craft industries.