Paper Type
ERF
Abstract
AI-powered streamers are increasingly appearing on live streaming platforms, producing content and interacting with audiences in ways that resemble human streamers. However, prior AI research offers mixed insights, and we still lack a clear understanding of how AI streamers shape viewers’ community perceptions in live streaming environments. Drawing on sense of community and uncanny valley theory, we develop a conceptual framework that explains how two AI-specific characteristics, anthropomorphism and AI efficacy, jointly influence viewers’ community perceptions. We propose an inverted U relationship between anthropomorphism and sense of community, with the strongest negative impact occurring at near human levels. Perceived AI efficacy moderates this relationship by strengthening the positive effects at low to moderate anthropomorphism and buffering the negative effects in the near human zone. Our framework clarifies the dilemma of AI-mediated social interaction and highlights implications for value co-creation on the live streaming platform.
Paper Number
1554
Recommended Citation
Liu, Yuqian and Deng, Chaoqun, "When AI Streamers Build or Break Live Streaming Communities: The Roles of Anthropomorphism, AI Efficacy, and Sense of Community" (2026). AMCIS 2026 Proceedings. 8.
https://aisel.aisnet.org/amcis2026/sig_dspe/sig_dspe/8
When AI Streamers Build or Break Live Streaming Communities: The Roles of Anthropomorphism, AI Efficacy, and Sense of Community
AI-powered streamers are increasingly appearing on live streaming platforms, producing content and interacting with audiences in ways that resemble human streamers. However, prior AI research offers mixed insights, and we still lack a clear understanding of how AI streamers shape viewers’ community perceptions in live streaming environments. Drawing on sense of community and uncanny valley theory, we develop a conceptual framework that explains how two AI-specific characteristics, anthropomorphism and AI efficacy, jointly influence viewers’ community perceptions. We propose an inverted U relationship between anthropomorphism and sense of community, with the strongest negative impact occurring at near human levels. Perceived AI efficacy moderates this relationship by strengthening the positive effects at low to moderate anthropomorphism and buffering the negative effects in the near human zone. Our framework clarifies the dilemma of AI-mediated social interaction and highlights implications for value co-creation on the live streaming platform.
Comments
SIG DSPI