Examining the Effects of Human-Likeness of Avatars on Emotion Perception and Emotion Elicitation
🤖 AI Summary
This study investigates how anthropomorphism of virtual avatars modulates perception and induction of Ekman’s six basic emotions, examining underlying mechanisms in nonverbal affective communication and potential discomfort (e.g., the uncanny valley effect). Using an experimental psychology paradigm, avatar anthropomorphism was systematically manipulated across three levels (low/medium/high), with multimodal assessment including emotion recognition accuracy, subjective affective ratings, and physiological measures. Results show that medium-anthropomorphic avatars with high “cuteness” (e.g., raccoon, shark) significantly enhance emotion recognition accuracy and elicit stronger positive affect; conversely, highly anthropomorphic avatars trigger negative affect, confirming the uncanny valley effect. The key contribution lies in demonstrating that emotional transmission efficacy is jointly moderated by cuteness and naturalistic facial dynamics—challenging the unidimensional similarity-based explanation of the uncanny valley. These findings provide empirical support and an integrated aesthetic–cognitive framework for designing empathic virtual agents.
Application Category
📝 Abstract
An increasing number of online interaction settings now provide the possibility to visually represent oneself via an animated avatar instead of a video stream. Benefits include protecting the communicator's privacy while still providing a means to express their individuality. In consequence, there has been a surge in means for avatar-based personalization, ranging from classic human representations to animals, food items, and more. However, using avatars also has drawbacks. Depending on the human-likeness of the avatar and the corresponding disparities between the avatar and the original expresser, avatars may elicit discomfort or even hinder effective nonverbal communication by distorting emotion perception. This study examines the relationship between the human-likeness of virtual avatars and emotion perception for Ekman's six "basic emotions". Research reveals that avatars with varying degrees of human-likeness have distinct effects on emotion perception. High human-likeness avatars, such as human avatars, tend to elicit more negative emotional responses from users, a phenomenon that is consistent with the concept of Uncanny Valley in aesthetics, which suggests that closely resembling humans can provoke negative emotional responses. Conversely, a raccoon avatar and a shark avatar, known as cuteness, which exhibit moderate human similarity in this study, demonstrate a positive influence on emotion perception. Our initial results suggest that the human-likeness of avatars is an important factor for emotion perception. The results from the follow-up study further suggest that the cuteness of avatars and their natural facial status may also play a significant role in emotion perception and elicitation. We discuss practical implications for strategically conveying specific human behavioral messages through avatars in multiple applications, such as business and counseling.
Problem
Research questions and friction points this paper is trying to address.
Investigates how avatar human-likeness affects emotion perception
Explores Uncanny Valley effect in high human-likeness avatars
Examines cuteness impact on emotion elicitation in avatars
Innovation
Methods, ideas, or system contributions that make the work stand out.
Avatars with varying human-likeness affect emotion perception
High human-likeness avatars elicit negative emotional responses
Cute avatars positively influence emotion perception
🔎 Similar Papers
