Impact of Artificial Intelligence Tools on Indian Students: Using UTAUT and Machine Learning Model

Abstract

The rapid advancement of Artificial Intelligence (AI) technologies has led to a transformation in higher education. AI tools provide academic support to students, available anytime, anywhere, to enhance their knowledge and skills. This study focuses on understanding the determinants of AI tools’ acceptance and use for academic support among students, influencing student satisfaction and academic performance in India. This work presents a quantitative investigation into the adoption of AI tools, including Chat-GPT, Google Gemini, Microsoft Co-Pilot, and related platforms, using the Unified Theory of Acceptance and Use of Technology (UTAUT). Data is collected from Kaggle, which contains 3,614 student records drawn from 1,246 institutions across 34 states in India. UTAUT constructs were operationalized as composite proxies from the available survey variables: Performance Expectancy, Effort Expectancy, Social Influence, and Facilitating Conditions (FC). The Actual Use Behaviour (USE) construct captured daily interaction intensity and contextual application diversity. A multi-layered Machine Learning (ML) model is employed, integrating Decision Tree, Logistic Regression, Random Forest, XGBoost, Gradient Boosting, and an ensemble Stacking Classifier. Class imbalance in the target variable addressed via SMOTE-ENN resampling. The final XGBoost model achieved al classification accuracy of 98.48% and an AUC score of 0.9986 and Cross-validated mean accuracy was 94.63% (SD = 0.0021). SHAP analysis provided post-hoc explainability, identifying USE as the dominant predictive feature in descending order of importance. The study contributes a theory-driven, methodologically transparent framework for analyzing AI adoption in higher education and offers practical implications for policy and institutional strategy, emphasizing the importance of structured support systems, digital competency development, and pedagogically aligned AI implementation. The paper concludes by emphasizing the need for a balanced approach to AI integration in education. Promoting AI literacy, implementing ethical AI policies, encouraging critical thinking exercises, and providing mental health support are recommended. By fostering responsible AI use, Indian educational institutions can maximize benefits while minimizing cognitive and psychological drawbacks.