Can chatbots represent? Examining the potential for symbolic representation in automated service agents

This study investigates citizens’ perceptions of automated service agents with human-like traits, within the context of public higher education. Drawing from representative bureaucracy and human-computer interaction literature, we assess the potential for symbolic representation by comparing chatbots to human agents. Using a conjoint experiment with visual demographic cues, we find a general preference for human agents over chatbots regardless of the demographic match. While there is limited evidence for symbolic representation generated by chatbots, citizens prefer matching-race service agents when they are human. This study reflects on the role of chatbots as service agents and suggests directions for future research.