After our conversation about computational aesthetics, something I kept thinking about that we didn’t discuss in much detail is the appraisal of human beauty. From several articles for the week, a common theme was how researchers, scientists, and artists grapple with how to measure aesthetics. However, a challenge is that aesthetics, what is deemed beautiful, its inherent subjectivity. While there are certain rules for what makes images aesthetic, like the golden ratio and rule of thirds, ultimately what is beautiful relies on the viewer. As described by Hoenig in his article about the definition of computational aesthetics, “Aesthetics is assumed always to be subjective, but aesthetics choices can reflect the opinion of either (a) one person, (b) a group of persons or (c), a normalized observer that represents some kind of universal aesthetic opinion” (Hoenig, p. 16). Moreover, the difficulties (and dangers) of how to measure aesthetics are further amplified in the discussion of human beauty.
Thus, I thought I would share two examples of the consequences of computational aesthetics of human beauty.
[1] First, around a year ago, I came across this website:
https://www.hownormalami.eu/. It’s described as an “interactive documentary” that teaches about different face analysis software and AI while simultaneously using your face to demonstrate how the software/AI works. While they go through several different AI metrics like predicted BMI and gender, the first one they present is about how “hot” you are on a scale of 1-10. For this beauty rating, they discuss how most of these algorithms are trained based on amassed ratings typically collected from university students. However, depending on where the ratings were collected, the AI may only be trained on the beauty norms from a specific culture. This can have a significant impact because companies will use beauty ratings as a part of their services. For example, Tinder will present matches of individuals with similar beauty ratings.
[2] In a similar vein, my second example is about augmented reality face filters on social media websites like Instagram and TikTok. For context, these social media sites have a suite of AR filters that allow individuals to change their appearance. However, there’s been a surge of negative reactions about how filters, especially ones that enhance beauty, can be especially damaging to a person’s self-perception. In fact, Forbes coined the term “Snapchat Dysmorphia” in relation to how constant usage of beauty filters causes individuals to become dissatisfied with their actual faces
https://www.forbes.com/sites/annahaines ... 6187a14eff. Furthermore, a recent empirical study shows that a primary reason people use these filters is to present their ideal, aspirational selves. However, using beauty filters for this reason also decreases self-acceptance and positive affect (Javornik et al., 2022).
Example of different beauty filters. Retrieved from
https://www.refinery29.com/en-gb/instag ... dysmorphia
Sources:
- Hoenig, F. (2005). Defining Computational Aesthetics (L. Neumann, M. Sbert, B. Gooch, & W. Purgathofer (editors, Eds.)
- Javornik, A., Marder, B., Barhorst, J. B., McLean, G., Rogers, Y., Marshall, P., & Warlop, L. (2022). ‘What lies behind the filter?’ Uncovering the motivations for using augmented reality (AR) face filters on social media and their effect on well-being. Computers in Human Behavior, 128, 107126