Airport Self-Service Technologies: Are We Fully Engaged? A Hybrid SEM-Machine Learning Approach

Ensuring the sustained use of self-service technologies (SSTs) is essential for companies investing in these innovations. This study investigates key factors shaping human attitudes toward SST adoption in airports by developing a conceptual model that examines the effects of perceived usefulness (PU), perceived ease of use (PEOU), and the need for human interaction (NHI) on utilitarian and hedonic attitudes. Structural Equation Modelling (SEM) is used to test the model, with age and gender assessed as moderating variables. Data were collected from 214 passengers at Muscat International Airport and enriched with interview insights. The model explains 76.2% of the variance in passengers’ intentions to use SSTs. PU and PEOU significantly influence both hedonic and utilitarian attitudes, while NHI shows no significant effect. Gender moderates the relationship between PU, PEOU, and utilitarian attitudes, while age moderates the link between hedonic attitudes and SST usage intention. To complement the SEM analysis and address behavioral complexities, five machine learning (ML) models—Decision Tree, Random Forest, Support Vector Machine, Artificial Neural Network, and Extreme Gradient Boosting—are employed to predict SST adoption. These models achieve an average prediction accuracy of approximately 96%. The integration of SEM and ML provides both explanatory depth and predictive strength, enhancing the understanding of behavioral drivers and supporting more informed SST implementation. Overall, this research offers practical implications for airport authorities and technology developers, emphasizing the importance of designing SSTs that balance functional efficiency with user engagement, while also considering demographic differences in technology acceptance. • Examines hedonic and utilitarian attitudes toward SSTs at airports • Gender and age moderate links between attitudes and SST intentions • Combines SEM and five ML models for behavioral prediction and insight • ML models reach 96% accuracy in forecasting SST usage behavior • Provides actionable guidance for SST design and airport service planning