This study investigates how curatorial attribution (AI versus human) influences audience perception across Narrative Coherence, legitimacy, and Aesthetic-Emotional Engagement. Utilizing a 2 × 2 factorial design (N = 281), participants evaluated identical artworks under varying provenance labels. Results from PLS-SEM indicate that while Narrative Coherence was not statistically significant, a marginal trend suggests a possible influence of attribution. The “Human” label significantly enhances perceived legitimacy and Aesthetic-Emotional Engagement. Notably, Signal Detection Theory (SDT) analysis indicates limited discrimination ability, suggesting that participants were generally unable to reliably distinguish between AI and human authored texts. However, Aesthetic-Emotional Engagement emerged as the primary significant mediator of Overall Evaluation. These findings demonstrate an attribution-based asymmetry, where evaluations appear to be shaped more by attribution cues than by accurate source identification. This research extends the Model of Aesthetic Experience, offering strategic implications for galleries regarding the integration of AI efficiency with human expertise.
Cahaya et al. (Thu,) studied this question.