A Hybrid SEM-Artificial Neural Network Study on Students’ Usage and Perceptions of ChatGPT: Exploring Academic Work Engagement

Abstract

The emergence of artificial intelligence (AI), particularly ChatGPT, has become widely used to aid students in educational tasks. This study investigated students' usage and perceptions of ChatGPT in academic work engagement using a revised Technology Acceptance Model (TAM) and Theory of Planned Behavior (TPB). An online questionnaire with 55 items was distributed, collecting 315 responses. Eleven latent variables were examined: technology readiness, user engagement, perceived ease of use, perceived usefulness, social influence, academic work engagement, attitude towards using ChatGPT, self-efficacy, response quality, intention to use, and actual use of ChatGPT. Structural Equation Modeling (SEM) revealed that technology readiness positively correlated with user engagement. User engagement significantly influenced social influence, affecting academic work engagement. Intention to use was significantly influenced by attitude towards using, response quality, and was related to the actual use of ChatGPT. Integrating Artificial Neural Networks (ANN) indicated the intention to use it as the most influential factor. The findings provide insights into the factors influencing ChatGPT adoption in academic environments and enhance the platform's performance.

Save as a Favourite

Author Information
Klint Allen Mariñas, Mapua University, Philippines
Charmine Sheena Saflor, De La Salle University, Philippines
Jasper Von Chico, Occidental Mindoro State College, Philippines
Hannah Maureen Manzano, Occidental Mindoro State College, Philippines
Monica Shane Tanglao, Occidental Mindoro State College, Philippines