Shifting Patterns Across Time: Evolving Associations Between AI Addiction and Dehumanization in a Three-Wave Longitudinal Study

The longitudinal association between AI addiction (AIA) and dehumanization, as well as its underlying mechanisms, remains unclear. This study examined the associations between AIA and dehumanization across three waves and tested the mediating role of information overload. A total of 492 university students from a province in China completed all three waves (M age = 18.74, SD = 1.01; 53.86% female). Restricted cubic spline analyses showed a nonlinear association between AIA and dehumanization at Time 1, but not at later waves. Random-intercept cross-lagged panel models showed no significant cross-lagged effects from Time 1 to Time 2. However, from Time 2 to Time 3, AIA and dehumanization showed significant reciprocal predictive effects. Information overload did not mediate these associations. These findings highlight the importance of examining temporal dynamics in understanding the psychological consequences of AI addiction. The association between AIA and dehumanization may become reciprocal over time.