Avoiding embarrassment online: Response to and inferences about chatbots when purchases activate self-presentation concerns

Abstract

We explore how self-presentation concerns and the desire to avoid embarrassment impact two distinct types of interactions consumers have with chatbots: interactions when a chatbot's identity is (1) not disclosed and therefore ambiguous or (2) disclosed. We propose that consumers feel less embarrassed with a chatbot than a human service agent in purchase contexts where self-presentation concerns are active because consumers ascribe less mind to chatbots. Therefore, when a chat agent's identity is ambiguous, consumers with greater self-presentation concerns are more likely to infer that an agent is human because this judgment allows consumers to proactively protect themselves from potential embarrassment in the event they are interacting with a human. We further show that when agent identity is clearly disclosed, consumers respond more positively to chatbots than human agents. However, this effect is contingent on the extent to which the chatbot is imbued with human characteristics: Anthropomorphizing chatbots leads consumers with higher self-presentation concerns to ascribe more mind to even clearly identified chatbots, resulting in a more negative consumer response.