Perception of Human Characteristics in an Ai Chatbot Acting as On-the-Job Coach

Abstract

With the increasingly widespread use of large language models (LLMs) in various applications, understanding the way such models are perceived by their human interlocutors has become more relevant. We explored the possibility of an AI workplace motivational coach using our chatbot HARRi, and whether such a coach would be perceived to have anthropomorphic characteristics such as emotions and empathy. Participants were told to converse with the bot as if they were logging in to work on time or late and were primed by instructions to believe they were either speaking to a chatbot or a human manager. Following the conversation with the chatbot, participants were asked to complete survey questions derived from the Individual Differences in Anthropomorphism Questionnaire (Waytz, Cacioppo, & Epley, 2010). Results indicated that with the exception of the human prime being rated as slightly more friendly, the bot and human prime made little difference in ratings of anthropomorphism. The implications of these findings are discussed.

Keywords: anthropomorphism, chatbot, IDAQ, Inspira