Discussing the Aesthetic Emotion of Artworks by AI and Human Artists with the Mediating Variable of Aesthetic Fluency

At the current stage, there is almost no aesthetic emotion research that directly explores the relationship between artistic creation and expertise in AI. Therefore, this study attempts to fill this gap and uses the Aesthetic Fluency Scale compiled by (Smith Empirical Stud. Arts 24(2)229-242, 2006) and the Geneva EmotionWheel (3.0) created by the SwissCentre for Affective Sciences to confirm that viewers with richer artistic expertise (the tested expert group, higher levels of aesthetic fluency) will be more interested in AI creation, winning more diverse and higher levels of aesthetic emotion. Besides, Viewerswith relatively inadequate artistic expertise (tested novice group, lower levels of aesthetic fluency) will feel more confused and less interested in AI creation, winning a relatively homogeneous and lower aesthetic emotion. Though some of the AI artworks involved in the experiments are already of a fairly high technical level and bear a close resemblance to abstract artworks by human artists in appearance, the results of this study confirm the previous conjecture due to the strong aesthetic-emotional connection between the viewer's level of expertise and the artworks. After collecting and analyzing the experimental data with SPSS software, this study has concluded that there are still significant differences in the emotional perception level between artworks created by AI and those by humans based on the different expertise and different aesthetic experience and emotional perception of artworks. Viewers with more expertise (the tested expert group) always have richer and higher levels of emotional perception in face of artworks by AI and human artists than viewers with less expertise (the tested novice group). Also, viewers with a higher level of expertise (the tested expert group) are more likely to give positive emotional evaluations of the AI artworks and are more tolerant of the AI artist's identity than viewers with a lower level of expertise (the tested novice group). This finding has some implications for the study of AI art creation.