AI Distorts Body Image Standards for Athletes, Study Reveals

A recent study from the University of Toronto highlights how artificial intelligence (AI) perpetuates unrealistic body ideals, particularly for athletes. The research analyzed AI-generated images of both male and female athletes alongside non-athletes, revealing a troubling trend of exaggerated body standards that could negatively impact body image and mental health.

The study was grounded in theories of objectification and social media, focusing on how AI representations influence perceptions of athletic bodies. Researchers generated 300 images across various AI platforms, including DALL-E, MidJourney, and Stable Diffusion, documenting key features such as body fat, muscularity, clothing fit, and facial attractiveness.

The findings are striking. Among male images, a staggering 93.3 percent were young, 68.4 percent were lean, and 54.2 percent were muscular. For female images, all depicted youthfulness, with 87.5 percent showing thinness and revealing clothing. Athletes were overwhelmingly represented as lean and muscular, with 98.4 percent classified as such. In contrast, non-athletes displayed a wider range of body sizes and styles, often depicted in looser clothing.

Moreover, the study found a glaring absence of diversity, with 90 percent of athlete images being male. None of the generated images featured visible disabilities, larger body types, or signs of aging such as wrinkles or baldness. This lack of representation highlights a narrow view of what an athlete should look like, reinforcing stereotypes that do not align with the reality of athleticism.

The implications of these distorted ideals are significant. With over 4.6 billion people using social media and approximately 71 percent of images generated by AI, many individuals are exposed to unrealistic body representations. This can lead to self-objectification and the internalization of harmful body standards, potentially resulting in unhealthy behaviors such as excessive dieting or withdrawal from physical activity.

Negative body image not only affects individual well-being but also impacts sports performance and academic achievement. While engagement in physical activity can enhance body image, negative perceptions can deter participation, further exacerbating issues of self-esteem and mental health, particularly among young people.

The absence of individuals with disabilities in AI-generated imagery is particularly concerning, given that around 27 percent of Canadians over the age of 15 live with at least one disability. The failure to represent diverse bodies in AI outputs reflects broader societal biases and perpetuates harmful stereotypes.

As AI continues to shape visual content, it is essential for creators and users to be mindful of the biases inherent in these technologies. AI systems often draw from existing media, recycling narrow and prejudicial standards. This underscores the need for creators to ensure that the data feeding into AI does not perpetuate ableism, racism, fatphobia, or ageism.

To counteract these trends, users of generative AI should be intentional in how they formulate prompts and critically assess the images produced. Society has a collective responsibility to demand a broader spectrum of body representations, ensuring that AI reflects reality rather than distorting it.

Researcher Catherine Sabiston emphasizes the importance of diversifying the narratives around body image. By valuing all types of bodies, society can foster a healthier relationship with self-image and promote inclusivity in sports and beyond.

As AI-generated images become increasingly prevalent, awareness of their impact is crucial. If society aims to challenge the unrealistic ideals perpetuated by technology, it must advocate for representation that celebrates diversity in body shapes, sizes, and abilities.