AI Content, User Fatigue, and Churn on TikTok: Testing an Integrated Stressor Model

The rapid proliferation of artificial intelligence–generated content (AIGC) on short‐form video platforms raises fundamental questions regarding its effects on users’ psychological well‐being and platform retention. Despite growing scholarly interest in social media fatigue and discontinuance, no empirical study has simultaneously examined AI‐related content overload and perceived content inauthenticity as co‐occurring stressors within a unified theoretical framework. Moreover, their differential influence on user strain and exit intention has not been systematically assessed in the context of short‐form videos. This study addresses these gaps by investigating how perceived AI content overload and perceived content inauthenticity shape social media fatigue, user dissatisfaction, and discontinuance intention among TikTok users in Vietnam. Drawing on an integrated stressor–strain–outcome (SSO) and push–pull–mooring (PPM) framework, a quantitative cross‐sectional survey was administered to 495 Vietnamese TikTok users, and the proposed model was evaluated using partial least squares structural equation modeling (PLS‐SEM). The results indicated that perceived AI content overload constituted the dominant stressor, significantly predicting both social media fatigue and user dissatisfaction, while perceived content inauthenticity produced comparatively modest yet statistically significant strain effects. Social media fatigue, but not dissatisfaction, significantly predicted discontinuance intention, and fatigue partially mediated the overload–discontinuance relationship. Habitual TikTok use emerged as the most powerful mooring factor, substantially suppressing users’ discontinuance intention. Collectively, the model accounted for 63.0% of the variance in discontinuance intention. These findings extend the SSO and PPM frameworks to AI‐mediated short‐form video contexts and yield actionable implications for platform designers, content governance practitioners, and regulators seeking to mitigate AIGC‐driven user retention.