Human-like or robot-like AI chatbots? Consumer’s preference of the anthropomorphism of AI chatbots under the advance selling

Purpose This research applies construal level theory (CLT) to explain the mechanism of advance selling consumers’ preference to the anthropomorphism level (human-like or robot-like) of artificial intelligence (AI) chatbots. Design/methodology/approach This research conducted three studies. Study 1 explored consumers' preferences for the anthropomorphism level of AI chatbots under spot selling and advanced selling conditions. Study 2 examined the mediating effect of psychological distance and ruled out the mediating role of uncertainty. Study 3 further investigated the boundary conditions of these effects under different levels of consumer expected effort. Findings According to the results of the studies, we have proved that consumers prefer to communicate with human-like AI chatbots in advance selling, and this is caused by the increased temporal psychological distance. However, this effect would be weakened by the significant consumer’s expected efforts. Originality/value This research extends the application of CLT to explain consumers’ different attitudes towards human-like and robot-like AI chatbots. In this study, the fluency of thought generated by expected product effort is proposed as the regulatory variable of temporal psychological distance with a time constraint, which provides a valuable boundary effect for CLT in consumer research and enriches the theoretical development of CLT within this field.