Exploring the Factors Influencing Generative AI Integration for Managing Self-Directed Learning: A Case Study on Indonesian Students’ Adoption of Google

The rapid incorporation of generative artificial intelligence (AI) into education has profoundly influenced self-directed learning (SDL). However, despite its rising importance, empirical investigations into the drivers of students’ adoption and continued use of such technologies, specifically Google Gemini, are still scarce. This study explores how Indonesian university students engage with Google Gemini for SDL, employing the UTAUT2 framework. Data were collected through a 36-item survey covering nine constructs, including Effort Expectancy (EE), Facilitating Conditions (FC), Habit (HA), Hedonic Motivation (HM), Price Value (PV), Performance Expectancy (PE), Social Influence (SI), Behavioral Intention (BI), and Google Gemini Use Behavior (GUB). A cross-sectional online survey was conducted among students from multiple levels and academic fields, yielding 514 valid responses. Partial Least Squares Structural Equation Modeling (PLS-SEM) was applied to test the hypothesized relationships among the UTAUT2 constructs, behavioral intentions, and usage behaviors. Findings revealed that most proposed pathways were supported, with SI exerting the strongest influence on BI, followed by PE, EE, and HA. BI, in turn, proved to be a robust predictor of GUB. Conversely, FC, HM, and PV were not significant predictors of intention. Both the measurement and structural models demonstrated satisfactory reliability and validity. The findings suggest that adoption of Google Gemini for SDL is shaped primarily by social endorsement and perceived utility, reinforced by ease of use and established routines, whereas infrastructural support and enjoyment play comparatively limited roles. These insights highlight the importance of peer norms, clear value propositions, and user-centered design in efforts to mainstream AI-assisted learning in higher education. Practical recommendations for educators, developers, and policymakers are outlined, alongside directions for future research on long-term engagement and learning outcomes.