Narrative Reconstruction and Aesthetic Shifts in AIGC Generated Imagery: Pathways for Building an Autonomous Chinese Knowledge System of Audiovisual Arts in the Age of Artificial Intelligence

Abstract

Generative artificial intelligence, commonly referred to as AIGC, produces images and videos through text prompts and is transforming the production mechanisms, narrative structures, and aesthetic standards of audiovisual art. AIGC imagery relies on data training and probabilistic sampling to create visually convincing results. Its narratives are often constructed retrospectively on the basis of generated content, while authorship is shifting from direct control of the camera to the design and refinement of prompts. Drawing on observations from creative practice, this study examines recurring problems in AIGC imagery, including character inconsistency, the breakdown of spatial continuity, and the inaccurate representation of cultural symbols. It further argues that an autonomous Chinese knowledge system of audiovisual arts should be developed through coordinated efforts at the conceptual, methodological, and cultural levels. The study suggests that AIGC, as a generative medium, expands the possibilities of artistic expression while also creating risks of aesthetic homogenization and cultural misinterpretation. Developing an appropriate theoretical response to these changes is therefore essential to the innovation of audiovisual arts in the age of artificial intelligence.