The Impact of Human–Chatbot Interaction on Human–Human Interaction: A Substitution or Complementary Effect

Abstract

Chatbots have been used in various contexts, enabling users to continually increase their trust in human-chatbot interaction (HCI). However, salient theoretical research and risk supervision remain insufficient. It is thus unclear whether HCI generates a substitution effect that threatens human–human interaction (HHI) or a complementary effect that facilitates HHI. Grounded in interdependence theory and trust transfer theory, this study proposes a theoretical model that explains how users’ established trust in HCI can affect their interpersonal communication intention. This theoretical model has been tested in both the task context and the social context. The results show that the effect of HCI on HHI varies in different contexts. In the task context, users prefer to interact only with the better service provider. There is thus a substitution effect of HCI on HHI. Users’ established trust in HCI makes them perceive that chatbot service providers offer more value than human service providers, thereby reducing their intention to interact with the latter. In the social context, users are open to interacting with multiple social partners instead of only the better partner. Hence, HCI has a complementary effect on HHI. Users’ established trust in HCI can also be extended to category-based trust in interaction with other unfamiliar social partners (unfamiliar people), thereby enhancing their intention to interact with unfamiliar people. Accordingly, these findings can enable the formulation of more targeted measures that mitigate the potential risks of HCI for HHI and promote future human-human relationships.