The Visual Language Reconfiguration Path of Generative Artificial Intelligence in Digital Art Creation

Abstract

Generative artificial intelligence (AI) has gradually entered the image generation, form deduction, style transfer and visual narrative
of digital art from creative assistant tools, which has changed the creation method with manual operation of software, preset visual elements
and linear production process as the core. Its influence does not stop at improving the efficiency of image production, but acts on the deep
structure of visual symbol formation, form organization, style expression and author decision-making. The generative model can combine visual elements in the latent representation space according to semantic instructions, so as to form a new transformation relationship among text
concepts, image features and style parameters, which also brings problems such as visual homogeneity, out-of-control form, semantic drift and
weakening of author subjectivity.