Understanding individuals' experience with virtual characters from the perspective of perceived authenticity

• Main Author: Huang, Junru
• Other Authors: Jung Younbo
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Social Sciences
• Online Access: https://hdl.handle.net/10356/175025
• Institution: Nanyang Technological University

Conventionally, human-controlled and machine-controlled virtual characters are studied separately under different theoretical frameworks based on the ontological nature of the particular virtual character. However, in recent years, technological advancement has made the boundaries between human and machine agency increasingly blurred. This dissertation proposes a theoretical framework that can explain how various virtual characters, regardless of their ontological agency, can be treated as unique social actors with a focus on perceived authenticity. Specifically, drawing on the authenticity model in computer-mediated communication and a typology of virtual characters, a multi-layered model of perceived authenticity of source is proposed to demonstrate how virtual characters do not have to be perceived as humans and yet can be perceived as authentic to their human interactants. Two experimental studies in the context of e-commerce live streaming were conducted to empirically test the proposed muli-layered model of perceived authenticity of source. Study 1 (N=104) employed a 2 (appearance: human-like v.s. machine-like) by 2 (claimed agency: human v.s. AI) between- subjects design and focuses on the mechanism of perceived authenticity via the first layer (i.e., perceived authenticity of claimed agency). Study 2 (N=310) employed a 2 (appearance: human-like v.s. machine-like) by 3 (claimed agency: human v.s. AI v.s. no revealment) by 2 (narrative type: real-life-oriented v.s. fiction-oriented) factorial design and investigated perceived authenticity of source on both layers (i.e., claimed agency and represented identity) and how they together influence perceived authenticity of source and subsequent affective outcomes and behavioral intentions. The results of the two studies provided preliminary evidence to support the proposed model. Implications, limitations, and directions for future research are discussed.