Transparency, Control, Diversity, and Responsibility:The Multidimensional Moderating Effects of Algorithmic Accountability on AI Service Avoidance

This study examines how algorithmic accountability dimensions moderate the relationships between perceived AI risks, expected service benefits, and AI service avoidance behavior. Drawing on privacy calculus theory, we propose that transparency, control, diversity, and responsibility function as governance mechanisms that buffer risk effects while amplifying benefit effects. Hierarchical regression analysis of data collected from 543 South Koreans reveals that perceived AI risks significantly increase service avoidance, while expected benefits reduce it. All four accountability dimensions moderate these relationships as predicted. Algorithmic control demonstrates the strongest buffering effect on risk-avoidance relationships, while algorithmic responsibility shows the greatest amplification of benefit effects. These findings extend privacy calculus theory by demonstrating that governance mechanisms fundamentally reshape the cost-benefit calculations underlying AI adoption decisions. The results provide evidence-based guidance for AI service providers and policymakers, suggesting that comprehensive accountability approaches addressing all four dimensions are most effective in promoting responsible AI adoption.