ARTIFICIAL INTELLIGENCE INTERACTION AND COMPUTER SELF-EFFICACY: A STRUCTURAL EQUATION MODELING APPROACH USING PLS-SEM

Abstract

The rapid expansion of artificial intelligence technologies has transformed the way individuals interact with digital systems, generating new forms of technological engagement and cognitive adaptation. Understanding the factors that influence user interaction with intelligent systems has therefore become an important area of research in information systems and digital behavior studies. The present study examines the structural relationships between artificial intelligence interaction, technology usage, computer self-efficacy, and user satisfaction through a structural equation modeling approach using Partial Least Squares (PLS-SEM).

A quantitative, cross-sectional design was implemented using simulated data from a sample of 400 users with experience interacting with digital technologies and artificial intelligence platforms. The measurement model was specified using reflective indicators measured through a five-point Likert scale. The structural model was estimated using the SmartPLS algorithm and evaluated through reliability, convergent validity, and predictive capacity indicators, including composite reliability, average variance extracted, path coefficients, and coefficients of determination.

The results indicate that artificial intelligence significantly influences technology usage, which in turn contributes to the development of computer self-efficacy. Computer self-efficacy emerged as the strongest predictor of user satisfaction, highlighting the central role of perceived technological competence in shaping positive experiences with digital systems. The model also demonstrates that computer self-efficacy acts as a mediating mechanism between technological interaction and satisfaction outcomes. The explanatory power of the model shows moderate to substantial levels of variance explained in the endogenous constructs.

These findings suggest that the successful integration of artificial intelligence technologies in digital environments depends not only on technological performance but also on users’ confidence in their ability to interact with intelligent systems. Strengthening technological competence and digital literacy may therefore enhance user satisfaction and facilitate the adoption of AI-based technologies.