Effect of ChatGPT's answering style on users' acceptance in a trip planning context

Abstract

This research examines the effect of the answering style of ChatGPT on users' acceptance in a trip planning context, where ChatGPT recommends a list of destinations. Focusing on two style factors (information structure: destinations are listed vs. explained; communication style: opening and ending remarks are present vs. absent), we examined how travelers' acceptance of ChatGPT's recommendation varied with its answering style based on the “Computers are Social Actors” paradigm, via three experiments. While the relationship between communication style and travelers' acceptance was not significant, that between information structure and acceptance was significant, travelers were more willing to visit the destinations recommended by ChatGPT when they were explained than when listed. Also, this research identified a mediator (perceived informativeness) and a moderator (travelers' familiarity with the destinations) of the relationship. As one of the earliest empirical studies regarding travelers' usage of ChatGPT, this study serves as a reference for future research.