Applying a modified technology acceptance model to explain higher education students’ usage of ChatGPT: A serial multiple mediation model with knowledge sharing as a moderator

Abstract

Generative artificial intelligence (AI), such as ChatGPT, has taken the world by storm, especially in the education sector, because of its capacity to produce responses that are contextually relevant and appear to imitate human language. This has increased concerns from both scholars and practitioners regarding the potential impacts of ChatGPT on students' learning. However, research on higher education students' adoption of ChatGPT is still scant. Drawing on the modified technology acceptance model (TAM) and a sample of 1389 higher education students recruited in 11 universities in Vietnam with a stratified random sampling approach, the findings of this study indicated that effort expectancy not only directly affected students' actual usage of ChatGPT, but also serially indirectly increased their actual use of ChatGPT through performance expectancy and intentions to use ChatGPT. Additionally, knowledge sharing was found to significantly increase higher education students’ transformation from having the intention to use ChatGPT to actual users of ChatGPT. The theoretical and managerial implications of this are discussed in this paper in order to gain benefits and manage the potential threats from this new technology.