Satiation of generative AI images

Abstract

Generative artificial intelligence (AI) has shown efficiency in creating novel images. However, limited studies have undertaken further questions, to what extent should generative AI created images be used, and do they surpass the effect of real ones? Based on hedonic adaptation theory, two experimental studies were conducted to determine the satiation effect of generative AI and real images. Study 1 found that generative AI images evoked a high level of inspiration in the beginning, which then steadily declined and showed the returning phase to the initial level. Study 2, which provided empirical evidence of the satiation effect, obtained identical results. However, mixed images showed lower inspirational levels in most repetition sets. Theoretical and practical implications are indicated.