Technology Paternalism and the Reactance Deficit - Are Our Natural Protective Mechanisms Failing Us?

With rapidly advancing artificial intelligence (AI) technology, the potential for paternalistic artificial agents grows. Autonomy restrictions have a detrimental impact on psychological well-being, regardless of the source of the restriction. The inherent opacity of AI is impeding the anticipation of autonomy-influencing factors. Hence, autonomy restrictions might go unnoticed thereby impeding our natural protective mechanisms: reactance reactions. To investigate the complex relationship between AI paternalism, autonomy perceptions, and reactance reactions we conducted an experimental, scenario-based, repeated measures mixed methods (2x5) online study (N = 137). The study took place in an e-commerce setting where participants were asked to buy a T-shirt from a web-shop, with varying levels of AI influence from a budget app. Participants experienced one of five levels of paternalistic influence (Control Condition without AI influence, Nudging Condition, Reversible Restriction, Non-reversible Restriction, Covert Restriction) and provided feedback twice: immediately after interaction and after receiving information about AI-supported decision-making and its paternalistic potential. The study's findings provide evidence for the presence of a reactance deficit, showing that individuals may have difficulty activating their natural protective mechanisms when confronted with highly paternalistic AI systems.