essay · 2026-09-16

Visual Folk Narratives in Generative Art

Research indicates that generative AI can lower the visual content production threshold while sparking widespread public engagement in artistic expression, yet a significant gap remains between this technological potential and practical application regarding folk story revitalization. The core issue is how to effectively leverage these tools for innovative transformation without losing cultural meaning, which current systems often fail due to superficial pattern-matching rather than semantic depth. A specific solution involves using an inference stage taxonomy-guided prompt framework that structures generation through three distinct levels: Level 1 representing objective visual geometry and Level 2 representing the specific cultural concept or object. By anchoring generation in this symbolic depth, researchers ensure the resulting output carries mythological connotation (Level 3) rather than merely replicating aesthetic trends. This matters because it transforms how we cultivate proactive anticipation and comprehension of artworks' logic and context within generative art methodologies. The framework demonstrates that accurate semantic control layers are essential for moving beyond

Sources

• An inference stage taxonomy-guided prompt framework for culturally aware generation of indigenous motifs | Discover Artificial Intelligence | Springer Nature Li
• Mechanism of public behavioral intention to use generative AI for folk story image co-creation | npj Heritage Science
• The Art of Generative Narrativity - arXiv.org