Artificial intelligence (AI) appearance technologies, including beauty filters, augmented reality (AR) modification and generative avatars, have individualized digital self-presentation. Unlike traditional media, these technologies create self-referential appearance ideals that create novel forms of self-discrepancy and appearance anxiety. A systematic review was conducted in peer-reviewed databases for empirical studies published from 2021 to 2026 that studied self-representation, self-discrepancy, appearance distress, and protective psychological constructs within this context of AI, following PRISMA 2020 guidelines. Emergent patterns were identified by using the narrative thematic synthesis method across eligible studies. A total of eight empirical studies were found that fulfilled the inclusion criteria. The themes that emerged were as follows: (1) AI-enhanced self as a novel appearance ideal of the self; (2) self-discrepancy as an operative mechanism between the appearance of idealized and non-idealized selves and psychological distress; (3) contextually related and protective moderators; and (4) representation modality—between aestheticizing versus non-aestheticizing appearance modalities and psychological distress. The findings from the synthesis showed that self-discrepancy, social self-comparison, and self-objectification are the three main routes for appearance anxiety. Theoretically, these non-idealizing avatars, high self-similarity, and goal-oriented use contexts were considered as protective buffers but were not consistently quantified. Most importantly, there was no empirical study that directly assessed positive body image constructs or body appreciation. The current research mainly focuses on the pathology aspect of AI appearance technology, and positive body image is not investigated. Empirical studies should be conducted in the future with the clear aim of examining if constructive, nonidealized relationships with the technologically enhanced body can foster psychological resilience, and if so, how.
No takes yet. Share an insight, caveat, or question.
Soman et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: