Social Presence of Generative AI Conversational Agents and Users' Willingness to Communicate: The Mediating Role of Technology Trust

This study examines whether the social presence of generative AI conversational agents is associated with users' willingness to communicate and whether technology trust accounts for part of this relationship.The study treats interaction with a conversational agent as an informatics-enabled service encounter in which users evaluate both relational cues and the dependability of the underlying technology.Data were obtained through an online survey of 400 adult users in China who had interacted with a generative AI conversational agent during the previous three months.The measurement and structural models were estimated with covariance-based structural equation modeling in IBM SPSS Amos 26.The findings indicate that social presence has positive associations with technology trust and willingness to communicate.Technology trust is also positively associated with willingness to communicate.Bootstrap analysis shows a positive indirect association between social presence and communication willingness through technology trust, while the direct association remains significant.The pattern is consistent with partial statistical mediation.The study connects social presence research with trust in a specific technology and distinguishes communication willingness from broader adoption and continuance intentions.For digital service systems, the results suggest that socially responsive interaction can support communication, provided that conversational conduct is accompanied by reliable and helpful performance.Designers and service managers should therefore evaluate social cues together with response accuracy, consistency, and transparent handling of uncertainty.