ChatGPT Personalized and Humorous Recommendations

Abstract

This study examines the impact of personalized and humorous responses generated by ChatGPT on the acceptance of and satisfaction with travel recommendations. Studies 1A, 1B, and 1C consistently indicate that visit intention and recommendation satisfaction were significantly higher when ChatGPT provided personalized rather than humorous responses. Study 2 investigates the effects of response type on visit intention and finds that recommendation satisfaction was not significant when participants were informed that the recommendation agent was human. Study 3 indicates that participants' usage experience with ChatGPT moderated the effects and that participants' need for cognition influenced their acceptance of personalized responses. Study 4 demonstrates different personalization methods from various sources, including preference-matching and tailored recommendation styles.