A Quantitative Approach to Understanding Machine Learning Adoption for Cybersecurity in E-commerce through the UTAUT Model

Abstract

This study quantitatively investigates the adoption of machine learning (ML) for cybersecurity in e-commerce through the Unified Theory of Acceptance and Use of Technology (UTAUT) framework. Using the open-access Mendeley “UTAUT, Combining PLS-SEM and MLR” dataset and the OECD/BCG/INSEAD “Adoption of AI in Firms” dataset, the research employs Partial Least Squares Structural Equation Modeling (PLS-SEM) and Fractional Logit Regression to provide complementary analytical perspectives on adoption behavior. Results reveal that Performance Expectancy and Effort Expectancy are the strongest predictors of Behavioral Intention, while Trust and Regulatory Compliance Pressure positively influence ML adoption, and Perceived Risk exerts a negative effect. Behavioral Intention significantly mediates actual ML use, demonstrating that perceived usefulness and ease of implementation drive organizational uptake. Theoretically, this study extends the UTAUT model by integrating trust, perceived risk, and regulatory compliance as contextual and institutional determinants of adoption. This integration enhances the model’s explanatory power for technology acceptance in high-risk, security-sensitive environments. Practically, the findings offer actionable insights for policymakers, cybersecurity managers, and e-commerce firms: strengthening ML transparency, aligning systems with regulatory frameworks, and implementing trust-building and user training programs can sustain adoption and improve cybersecurity resilience. By combining rigorous empirical modeling with contextual extensions of UTAUT, this research contributes to a deeper understanding of how behavioral, technical, and governance factors jointly shape ML adoption for secure digital commerce.