Public acceptance of LLM-driven healthcare chatbots in China: An empirical study

Abstract

LLM-driven healthcare chatbots for preliminary medical consultation are a promising innovation to improve healthcare accessibility and efficiency. However, public acceptance of this technology in the Chinese context, especially the impact of users’ previous experience with relevant technologies on user behavior, remains underexplored. To address this gap, we extended the classical Unified Theory of Acceptance and Use of Technology (UTAUT) framework by examining the moderating effects of users’ previous experience with telemedicine and large language models (LLMs). Using a scenariobased survey, we collected 502 valid responses from general Chinese users and analyzed the data through Structural Equation Modelling (SEM). Our results demonstrated that performance expectancy, social influence, trust, and facilitating conditions were significant contributing factors, whereas effort expectancy was not, which contradicts previous literature. Moreover, users’ previous experience with LLMs exhibited significant moderating effects whereas previous experience with telemedicine didn’t. These findings contribute to the literature by suggesting that as LLMs become more widely adopted, users’ familiarity with them may enhance trust and, consequently, increase the general acceptance of LLM-driven healthcare chatbots.