The Challenge of Implicitness: Exploring AI’s Cultural Competence of Eastern Aesthetics through Poetic Painting Themes

Source of Publication

Proceedings of the 31st International Symposium on Electronic/Emerging Art (ISEA2026)

Conference Year

2026

Abstract

AI tools can support artistic creation, but translating culturally specific meanings into images remains difficult. This study examines how three large language models (LLMs) interpret implicit meanings in classical Chinese poetry through three poetic-painting themes associated with the Song dynasty Imperial Painting Academy. The research combines unguided prompt generation, human-guided iteration, and qualitative evaluation by three traditional-painting experts. The models used conventional painting vocabulary and sometimes developed inventive interpretations, but their prompts often favored literal imagery, repeated familiar solutions, or introduced details inconsistent with the verses. Some prompts also described sounds or movements that a static image could not directly represent. By examining both LLM-generated prompts and images rendered with a separate image generator, the study identifies difficulties in selecting which poetic elements to show, suggest, or leave unstated. The findings describe the tested workflow rather than a general limitation of all AI systems.

Document Type

Conference Proceeding

Publication Date

11-6-2026

This document is currently not available here.

Share

COinS