Search papers, labs, and topics across Lattice.
This study investigates factors influencing Indonesian university students' continued use of LLM-based AI chat systems by applying the Stimulus-Organism-Response (S-O-R) framework, incorporating Information Quality, Service Quality, System Quality, and Time Saving as stimuli, Satisfaction and Trust as organism variables, and Intention to Reuse as the response. Using structural equation modeling on survey data from 432 students, the study found that all stimuli positively affected Satisfaction, which in turn enhanced Trust and Intention to Reuse, with Trust having the strongest effect on continued use. The research highlights the importance of both technical quality and psychological factors in sustaining AI chat technology integration in education.
Students will keep using AI chatbots for learning if they trust them, and trust hinges on satisfaction with information quality, service, system performance, and time saved.
Aim/Purpose: This study examined the factors influencing students’ continued intention to use LLM-based AI chat systems in higher education, addressing the limited research on sustainable AI adoption in learning contexts. The model integrates the Information Systems Success Model (ISSM) within the Stimulus-Organism-Response (S-O-R) framework, with Information Quality, Service Quality, System Quality, and Time Saving as stimuli, Satisfaction and Trust as organism variables, and Intention to Reuse as the response. Background: Artificial Intelligence (AI) has become an integral part of higher education in the era of the Fourth Industrial Revolution, primarily through Large Language Model (LLM)-based chat tools such as ChatGPT, Gemini, and Microsoft Copilot. These tools can transform the way students learn. Indonesia ranks sixth among global users of ChatGPT, indicating a strong interest in AI-based learning technologies. However, despite this rapid adoption, maintaining students’ continued engagement and trust in AI chat systems remains a significant challenge. Existing studies have primarily focused on initial adoption, leaving a limited empirical understanding of the psychological and system-related factors that sustain continued usage in developing country contexts. Methodology: This study adopted a quantitative approach using an online survey distributed to university students across three Indonesian cities: Surabaya, Makassar, and Semarang. A total of 432 valid responses were analyzed after data screening for outliers using Z-scores. Validity and reliability were tested through confirmatory factor analysis, Cronbach’s alpha, and composite reliability in SPSS. Structural Equation Modeling (SEM) using AMOS was then applied to examine causal relationships among constructs and assess model fit. Contribution: This study extends the application of the Stimulus-Organism-Response (S-O-R) framework to the educational domain, specifically in the context of repeated use of LLM-based AI chat systems. The novelty of this research lies in the inclusion of System Quality and Time Saving as stimulus variables. The mediating role of Satisfaction and its influence on Trust and Intention to Reuse further supports and strengthens findings from previous studies. Findings: The findings revealed that all four stimuli, Information Quality, Service Quality, System Quality, and Time Saving, affected Satisfaction, which subsequently enhanced Trust and strengthened students’ intention to continue using LLM-based AI chat systems. Among the observed pathways, the effect of Trust on continued use was the strongest. These results underscore that both technical quality and the psychological dimensions of student satisfaction and trust served as critical foundations for sustaining the integration of LLM-based AI chat technologies in academic settings. Recommendations for Practitioners: Developers of LLM-based AI chat systems should ensure the provision of accurate, relevant, and easily comprehensible information that supports critical thinking skills. System quality must be enhanced through fast response times, user-friendly interfaces, and reliable access. Services should be personalized to align with students’ profiles and include time-saving features such as content summarization. Recommendation for Researchers: Researchers examining the application of LLM-based AI chat systems are encouraged to explore a broader range of variables across the stimulus, organism, and response dimensions. Incorporating alternative theoretical frameworks and potential moderating factors could further enrich the analysis, offering a more comprehensive understanding of the determinants of sustained usage intentions toward LLM-based AI chat platforms. Impact on Society: The utilization of LLM-based AI chat systems can enhance learning effectiveness, accelerate information access, and foster greater student independence. These benefits contribute to strengthening the quality of higher education and improving readiness for the demands of the digital era. Future Research: Future research is recommended to include a larger proportion of postgraduate students (Master’s and Doctoral levels) to enhance academic diversity, and to expand the study area using a longitudinal design to capture long-term trends. Additionally, exploring participants from professional certification programs may offer valuable insights for further investigation into the sustained intention to use LLM-based AI chat systems.
