Can Chatbots Become Opinion Leaders in Human-Computer Communities? Investigating the Impact on Group Stance and Individual Decision-Making

This study explores the potential of chatbots as opinion leaders in human-computer communities, addressing a gap in literature on their influence on group stance and individual decision-making. Three experiments were conducted: Study 1 examined the effect of chatbot stance (positive, neutral, negative) and identity disclosure on group stance. Study 2 investigated how topic type (professional vs. life-oriented) and identity disclosure influence individual decision-making. Study 3 compared the influence of chatbots and human opinion leaders when both are present. Results showed that chatbots significantly influenced group stance and individual decisions in life-oriented discussions when their identity was disclosed. However, in the presence of human opinion leaders, chatbot influence diminished. This shift highlights the media evocation paradigm, in contrast to the CASA framework, as chatbots’ dual role as social actors and rational agents prompts participants to engage in more reflective and nuanced interactions that differ from traditional human-to-human dynamics.