‘Please, Just Make Service Faster’: When Human-likeness in Chatbots Backfires
(A2023-113944)

AUTHORS

Lennart Seitz, Leuphana University Lueneburg; Sigrid Bekmeier-Feuerhahn, Leuphana University Lueneburg

ABSTRACT

The increasing use of chatbots in customer services has initiated a broad debate on how to make interactions with them more human-like. Although humanization can be beneficial, this research hypothesizes that social cues might backfire when harming perceived usefulness that is a key predictor of technology acceptance. More specifically, we assume that human-like response time delays violate users’ expectations towards a fast and efficient self-service thus reducing using intentions. In three experimental online-studies using video stimuli and real interactive chatbots, we find evidence for a negative effect of response time delays on using intention fully mediated by a detraction in perceived usefulness. We further find that user’s individual extent to anthropomorphize the chatbot moderates the effect, i.e., the effect is weaker for people who ascribe human-like traits to the agent. Theoretical and practical implications as well as future research avenues are discussed.