Human or AI? The Mediating Role of Perceived Risk and the Moderating Effect of Product Type on Purchase Intention in Livestreaming Commerce

서지정보

· 발행기관 : 한국생산성학회
· 수록지 정보 : 생산성연구: 국제융합학술지 / 39권 / 3호 / 5 ~ 42페이지
· 저자명 : 황의, 장정

초록

With the rapid integration of Artificial Intelligence (AI) technology into digital marketing, AI-powered virtual streamers (digital avatar streamers) have been widely implemented in livestreaming e-commerce contexts. Compared to traditional human streamers, virtual streamers may face limitations in emotional expression, authentic interaction, and personalized recommendation capabilities. These limitations can trigger consumers’ cognitive discomfort and perceived risk, which in turn may negatively influence their purchase decisions. However, existing research remains limited regarding how consumers perceive the risks associated with virtual streamers and the specific mechanisms by which streamer type affects purchase intention.
Guided by the Stimulus–Organism–Response (SOR) model and perceived risk theory, this study conducts two experiments to investigate the impact of different streamer types (human streamer, digital avatar streamer, and 3D anime-style virtual streamer) on consumers’ purchase intention. It further introduces product type (utilitarian vs. hedonic) as a moderating variable. Study 1, using a single-factor between-subjects design, finds that virtual streamers significantly increase perceived risk among consumers, thereby reducing their purchase intention. Perceived risk fully mediates the relationship between streamer type and purchase intention, revealing potential challenges that AI streamers face in achieving conversion effectiveness. Study 2 adopts a 2×2 between-subjects design and further reveals the moderating role of product type within the mediation process.
Specifically, when the product is hedonic, digital avatar streamers elicit higher perceived risk and lower purchase intention; by contrast, under utilitarian product conditions, this difference is not significant. This finding indicates that product type functions as a contextual boundary condition in the psychological mechanism through which AI streamers affect consumer behavior. Theoretically, this research enriches the literature on consumer behavior in the context of AI virtual streamers and extends the application of perceived risk theory.
Practically, it offers strategic insights for brands in selecting appropriate streamer types and matching product strategies, especially in precision marketing within AI-driven livestreaming commerce environments.

참고자료

· 없음