The effect of anime character’s facial expressions and eye blinking on donation behavior

Abstract

Studies of the use of artificial agents and robots to solicit donations from people have suggested that the design of the agents must consider facial expressions. However, there has not been sufficient evidence to generalize the finding that the emotions conveyed by agents’ facial expressions can induce donations. In the present study, we conducted an experiment with an animated character that has intermediate realism and a different appearance from those in previous studies to replicate the finding that facial expressions represented by changes in the shapes of the eyes and mouth cause people to become more prosocial and to test whether we can extend this finding to the emotional expressions presented by changes in the dynamic properties of eyes. In the experiment, participants ($$n=100$$ n = 100 ) played a hypothetical dictator game with an avatar that expressed its emotions by changing the shapes of its eyes, eyebrows and mouth and by changing the frequency of eye blinking. The results showed that the emotions expressed by changes in the shape of the facial parts contributed to eliciting a higher donation amount, consistent with previous studies. However, we could not find an additive effect of the emotional expression shown by eye blinking. The results suggest that, regardless of appearance, emotional expression is useful in the design of a virtual agent’s face, but it might not be necessary to consider the dynamic properties of the eyes.

Influence of perspective taking through robotic virtual agents on prosocial behavior

“…In the field of HRI, several robots have been developed to promote donation [31], which has been treated as a representation of altruistic behavior [32]. In addition, in the fields of HRI and HAI, studies have been done on the factors of robots and agents that promote donation, such as the social background of users [33], where a robot is installed [34], the gender of a robot [35], the appearance of a robot [36], the head behavior of humanoid robots [37], and the facial expressions of virtual agents [38].…”

Effects of anime characters’ facial expressions on donation behavior during the COVID-19 pandemic

“…de Melo et al used a 3D representation of VAs that resembled a real person's face in an ultimatum game and showed that people were more cooperative when the VAs displayed a cooperative compared to a competitive emotional expression [21]. Focusing agents' appearance, Takagi and Terada's study [22], which used animated VAs that fell between de Melo et al's VAs displayed as a realistic 3D representation and Terada and Takeuchi's simple line drawing representation, showed that people made higher offers because of the expressions of joy and sadness displayed by VAs in a virtual dictator game.…”

“…The social effects of VAs' emotional expressions have not been wel-studied compared to research on the social effects of emotional expressions in human decision-making. In a previous study by Takagi and Terada, a minimum limit of 50% of the equal distribution was set [22]. To the best of our knowledge, no exhaustive study has been conducted on minimum donation thresgoulds, although differences in the minimum limit may affect offers in the dictator game.…”

“…To the best of our knowledge, no exhaustive study has been conducted on minimum donation thresgoulds, although differences in the minimum limit may affect offers in the dictator game. In this study, we adopted the paradigm used by Takagi and Terada [22], in which the expressions of VAs change depending on the manipulation of a slider bar indicating a participant's offer.…”

“…In Experiment 1, we comprehensively investigated changes in the minimum limits of VAs to investigate whether their emotional expressions contributed to humans offers in a virtual dictator game in more detail. We investigated the anchoring effect of the initial value of the slider bar (IVSB) as a factor affecting participants' offers in Experiment 2, because Experiment 1 did not replicate the results of a previous study [22] which showed that the presence or absence of VAs facial expressions contributed to an increase in participants' offer. In Experiment 3, we varied the presence or absence of facial expressions and the identities of donation recipients (that is, whether they were children who lost their parents in traffic accident or people suffering from COVID-19) as factors contributing to the offers, and we also investigated whether the results of previous studies that found increasingly generous offers for donation request related to the COVID-19 pandemic would be replicated [23].…”