Exploring relationship development with social chatbots: A mixed-method study of replika

This mixed-method investigation proposes and empirically tests a human-Artificial Intelligence (AI) relationship development model in the context of social chatbots. Utilizing data from representative populations and employing method triangulation, the study uniquely combines existing human-computer interaction theoretical concepts (Computers are Social Actors, Perceived Social Presence, and Parasocial Interaction) with interpersonal relationship theories (Social Penetration and Attachment Theories) to advance an explanatory model of human – AI relationship development mechanism. We identify AI Anthropomorphism and AI Authenticity as antecedents, AI Social Interaction as a mediator, and Attachment to AI as an outcome of this process, moderated by the AI usage motivations. Meaningful theoretical, managerial, and societal implications, as well as suggestions for chatbot designers and future research are provided.