Research on the Influence of Aesthetic Features of Artificial Intelligence Generated Paintings on Perceived Value and Usage Intent

Rapid advances in text-to-image Artificial Intelligence Generated Content (AIGC) technologies are reshaping artistic creation and visual design workflows. Yet, the aesthetic quality of AIGC images varies substantially due to model limitations and heterogeneity in user prompts, underscoring the need for a systematic evaluation from the perspective of “operationalized aesthetic features–aesthetic appraisals–behavioral intention.” Drawing on the Stimulus–Organism–Response (S–O–R) framework and integrating visual variables, computational aesthetics, iconology, affective arousal, and perceived value theories, this study proposes an integrative model in which image aesthetics serve as stimuli—including surface-level formal features (composition, color, configurational form, and texture) and a deeper semantic feature (cultural-semantic content)—that shape organismic responses (aesthetic emotional experience, perceived emotional value, and perceived aesthetic value), ultimately predicting behavioral intention to use AI image-generation tools (BI). A stimulus-based online perceptual survey was conducted among participants with art/design backgrounds in Mainland China and South Korea, yielding 386 valid responses. Methodologically, Partial Least Squares Structural Equation Modeling (PLS-SEM) was employed to test the measurement and structural models, and a feed forward Artificial Neural Network (ANN) with strict 10-fold cross-validation was further used to assess out-of-sample predictive performance and variable importance.Findings indicate that composition, color, configurational form, and texture exert significant positive effects on aesthetic emotional experience and/or perceived aesthetic value, and also directly enhance BI. In contrast, cultural-semantic content significantly strengthens aesthetic emotional experience and perceived aesthetic value but shows no significant direct effect on BI. Mediation analyses reveal that perceived aesthetic value plays a pivotal mediating role between aesthetic features and BI, while aesthetic emotional experience influences BI indirectly through a serial pathway via perceived emotional value and perceived aesthetic value. The ANN-based cross-validation and sensitivity analyses provide predictive robustness for the proposed mechanisms and further identify the relative contribution ranking of inputs across endogenous predictions. Overall, this study advances the explanatory and predictive applicability of the S–O–R framework in the context of AIGC-driven aesthetic perception and technology adoption, and offers practical implications for aesthetic quality optimization and user-experience design of AI image-generation tools.