Factors influencing the acceptance and usage of ChatGPT as an emerging learning tool among higher education students in Bangladesh: a structural equation modeling

ChatGPT’s capability to provide immediate responses to student queries positions it as a potentially transformative educational tool. Nevertheless, its impact on Bangladeshi university students remains a subject of debate. This cross-sectional study examines ChatGPT usage among Bangladeshi university students and its determinants using the Technology Acceptance Model (TAM). Data from 729 students across five public and four private universities were analyzed via inferential statistics and Structural Equation Modeling (SEM). Results indicate perceived usefulness, ease of use, and perceived risk significantly influence ChatGPT adoption. Private university students’ usage was primarily driven by ease of use and perceived usefulness, while perceived risk and attitude showed no significant impact. In contrast, public university students’ usage was strongly influenced by perceived usefulness and existing knowledge, with perceived risk negatively affecting attitudes. Public university students perceived higher risks and lower ease of use than private peers. SEM highlighted ease of use as the strongest positive predictor in private institutions, while existing knowledge was more influential in public ones. The findings suggest structured training, awareness campaigns, and safety policies could mitigate risks and enhance ethical adoption. Public universities require targeted interventions to address risk perceptions, whereas private institutions benefit from emphasizing ChatGPT’s usability and academic value.