Body image refers to people’s mental picture of their physical appearance. Body image disturbance (BID) has become a widespread problem among young women and is implicated in low self-esteem and eating disorders. Media portrayals of unrealistically idealized images (e.g. super slender models and attractive celebrities) are accepted as the most powerful trigger and disseminator of young women’s BID.
However, with the introduction of deepfake technology which enables image synthesis between two different individuals, ordinary people can become part of the perfect image and virtually own the idealized bodies instead of being purely passive viewers. ZAO, an entertainment and apparel app enabled by deepfake technology, became the most downloaded app in China in August 2019. With the help of generative neural network, users can replace the celebrity’s face with their own as if they became the celebrities on the screen. Among all the source videos, one category hash-tagged #演女神# (#Act Goddess#, which means portraying extremely attractive female celebrities in Chinese) have received thousands of views and clicks. An emerging question is how young women will evaluate their appearance after exposure to the deepfaked self? This pilot study aims to investigate the impact of self-celebrity deepfaked video (SCDV) exposure on young women’s body image.
We proposed two competitive hypotheses. On the one hand, the SCDV which places one’s own face with the body of the celebrities should lead to cognitive dissonance and prime the social comparison process, which will result in self-depreciation. On the other hand, the SCDV will also visualize an 'attractive possible self' and create chances for fantasy of becoming more attractive, which may enhance viewers’ body satisfaction.
Through a between-subjects experiment with 128 young women aged between 18 and 31. We suggested that compared to purely celebrity video (PCV) exposure, SCDV exposure produced higher body satisfaction, F (1,125) =12.503, p=0.001, ηP2=0.091, and perception of appearance-improving potential served as the mediator (effect=0.058, SE=0.025, CI95% [0.017, 0.118]). Furthermore, participants’ body satisfaction was increased by SCDV exposure (t= -2.083, p=0.041) but decreased by PCV exposure (t=3.26, p=0.002).
Results indicate that compared to PCV exposure, SCSV exposure functions as an inspiration rather than a threat for young women's body image. One established explanation is that they will clearly perceive their potential to achieve a more attractive self through trying on different face shapes, hair styles, bodies and outfits that originally belonged to upward others.
This study is among the first to investigate the social impact of deepfake technology and the first to examine the effect of SCDV exposure on young women’s body image. It reveals the positive role of deepfake technology in entertainment and apparel app use. Considering the prevalence of BID among young women and commercial development of deepfake technology, such kind of mobile application may serve as a daily intervention technique for body dissatisfaction and low appearance self-esteem. Besides, this study also contributes the understanding of the self-enhancement effect and the notion of 'malleable self
' in the context of the avatar-based media.