Silenced by non-consensual deepfakes? Perceived threat and online self-censorship among female users: Mediating roles of perceived control, belief in a just world, and passive SNS use

This study examines whether perceived threat of “deepfakes” predicts online self-censorship among female users, testing perceived control, belief in a just world (BJW), and passive SNS use as mediators. An online survey of 423 young female users in South Korea assessed perceived deepfake threat, BJW, perceived locus of control, SNS engagement, and self-censorship behavior. Structural equation modeling was used to evaluate direct and indirect effects. Perceived deepfake threat was positively associated with self-censorship, fully mediated by perceived control and passive SNS usage. Higher threat perception predicted lower perceived control, which in turn increased self-censorship. Similarly, those fearing deepfakes reported increased passive SNS use, which was linked to greater self-censorship. While BJW was significantly associated with perceived control in the structural model, its indirect effect did not reach significance in bootstrapping. Female users who perceive a high deepfake threat do not self-censor automatically; rather, psychological mechanisms shape their responses. Threat perception might undermine an internal locus of control and discourages active online participation, ultimately leading to self-censorship. Findings highlight the chilling effects of AI-enabled image-based abuse. Interventions that enhance digital agency—such as privacy-enhancing tools, platform accountability, and belief in online justice—may mitigate defensive online behaviors.