From Intention to Habit in ChatGPT Use: An Examination Through Person-Centered Profiles

Abstract

This study examines the transition from intention to routine use in ChatGPT adoption in higher education from a person-centered perspective. A total of 351 valid responses from students and instructors were analyzed following data screening procedures. Although previous research on generative AI adoption has largely relied on variable-centered approaches such as UTAUT2 to explain average relationships among constructs, the present study argues that implementation challenges can be understood more effectively through user segments. Using k-means clustering on standardized UTAUT2 driver constructs, performance expectancy, effort expectancy, social influence, facilitating conditions, hedonic motivation, price value, and habit, three adoption profiles were identified: Capability-Ready Explorers, Routine Adopters, and Skeptical/Detached Users. To interpret differences among these groups more precisely, the study introduces two conversion indicators: the Capability-to-Habit Discrepancy Index (CHDI) and the Intention-to-Use Residual (IUR). Findings show that the largest group, Capability-Ready Explorers, demonstrates relatively high perceived ease of use and support but low habit and only moderate use frequency, indicating a gap between readiness and routinized practice. Variation across faculties and school years further suggests that ChatGPT adoption is shaped by disciplinary and developmental context. The discussion integrates UTAUT2, habit formation theory, and self-regulated learning to argue that successful institutional adoption depends not only on promoting positive attitudes, but also on designing repeatable, legitimate, and accountable routines of use. The study concludes that person-centered profiling offers a practical framework for understanding and supporting responsible AI integration in higher education.