Generative AI in advertising: Trust, Acceptance, And Rejection Dynamics Among Young Social-Media Users

Research Aim: This study investigates the determinants of generative AI rejection among teenage consumers in Central Java, specifically examining how declining trust and AI acceptance influence rejection behavior, with technology experience as a moderating variable. Design/Methods/Approach: A quantitative causal-comparative research design was employed utilizing Structural Equation Modeling based on Partial Least Squares (SEM-PLS) through SMARTPLS software. Data were collected from 289 teenage consumers (aged 18+) in Central Java who had exposure to AI-generated digital marketing content on social media platforms and marketplace applications. Purposive and snowball sampling techniques were used to ensure relevant respondent representation. Research Finding: The study revealed that declining trust in generative AI (β = 0.116, p < 0.05) significantly predicts consumer rejection behavior, whereas direct AI acceptance did not show a significant relationship with rejection (β = 0.032, p = 0.088). Technology experience demonstrated a significant moderating effect (β = 0.082) on the acceptance-rejection relationship. The model explained 68% of variance in rejection behavior (R² = 0.68), indicating substantial predictive power. Theoretical Contribution/Originality: This research extends existing technology acceptance models by integrating trust theory and behavioral intention-action gap framework, demonstrating that affective-cognitive factors (trust, authenticity, ethical considerations) outweigh utilitarian value in predicting AI rejection among teenagers. The findings bridge the gap between cross-sectional AI advertising effectiveness studies and long-term consumer psychological responses. Practitioner/Policy Implication: Organizations implementing generative AI in digital marketing should prioritize trust-building through transparent AI disclosure, ethical practices, and privacy protection. For tech-inexperienced consumer segments, enhanced AI literacy programs are essential to reduce rejection rates and foster sustainable brand relationships. Research Limitation: The study employed a cross-sectional design limiting temporal effect assessment. Geographic scope was restricted to Central Java, potentially limiting generalizability. Future research should incorporate longitudinal designs and additional variables such as AI literacy, innovation orientation, and negative technology experiences.