With the rising popularity of virtual reality, big data, and artificial intelligence (AI), social avatars have become important elements in user-generated content on social platforms. Social avatars can make use of network centrality to influence users' perceptions and attitudes toward products, lifestyles, and even trends, influencing their behavioral decisions and imitation intentions. This study explored the effect of the strength of social avatar centrality on users' intentions to imitate based on social network and parasocial interaction theory. The study determined the following: first, avatar centrality is positively related to consumers' imitation intent toward avatars; second, avatar centrality is positively related to self–AI connections and parasocial relationships; third, self–AI connection is positively related to parasocial relationships; fourth, self–AI connection and parasocial relationships are positively related to consumers' imitation intent toward social avatars. The findings of this study can enrich the application of social network theory in AI and provide important insights for social network platforms and brand promotion to strengthen the social relationships between social avatars and users and the social influence of avatars.