Skip Navigation
MarylandToday

Produced by the Office of Strategic Communications

Subscribe Now
Research

The Ugly Side of Beauty AI

How Tech Increasingly Pushes Narrow Ideas of Beauty

By Gregory Muraski

Beauty filters on smartphone

Photo by iStock

Apps that employ algorithms to analyze attractiveness—suggesting makeup strategies and even surgeries—often have racial bias cooked into the code, says a UMD researcher.

Our collective obsession with beauty is the baseline for industries ranging from cosmetics to fitness to various forms of surgery, and now it’s becoming integral to the development of artificial intelligence (AI). 

Increasingly popular “beauty filter” apps let users smooth out facial imperfections or tweak individual features, like bigger, doll-like eyes, a model-slim face or narrower nose. Services such as Qoves Studio and Face Plus Plus offer facial assessment tools that run on neural networks that mimic the operations of the human brain while using sample data of people’s faces to make recommendations—maybe to change makeup strategies, but also perhaps for a surgical intervention.

The pitfalls to AI-driven beauty scoring begin with the premise that while real-world beauty is in the eye of many different beholders with diverse viewpoints, app-scored beauty is subject to a single electronic gaze, said Lauren Rhue, assistant professor of information systems at the University of Maryland’s Robert H. Smith School of Business. 

“There are all these different cultural standards that have to do with beauty,” she said. “How can you train an algorithm to determine whether or not someone is beautiful?”

Rhue, whose research explores the economic and social implications of technology, said that though popular social platforms like TikTok, Instagram and Facebook have denied using such algorithms, the “recommendation algorithms”  embedded in these services to influence posts that you view often end up gauging attractiveness, whether intended or not.

“If you look at what Instagram wants, it's going to be essentially models, right? You're not going to see a lot of different types of facial features and expressions. And, and that's going to perpetuate this idea of beauty because… of the lack of diversity in what you see in Instagram, and what's extremely popular on Instagram,” Rhue said in a recent MIT Technology Review podcast.

Rhue acknowledges a certain entertainment value in beauty filter apps. “But our choice of beauty filters is definitely informed by the culture,” which is dominated by Eurocentric beauty standards that could exclude those from racial or ethnic minority populations.

In a working paper, Rhue is exploring how AI anchors subjective and objective predictions. She finds that women with lighter skin and hair consistently rate as more attractive than counterparts with darker skin and hair, indicating that filters that use facial detection likely have racial bias built in.

The serious implications of AI beauty apps are largely being overlooked by the tech community, she said.

“A lot of people say, ‘Oh, well, beauty is just not important because we're tech people and we're objective,’” she said. “But beauty is a huge industry. It has such an impact on people. And the idea that there isn't more research is really interesting to me.”

Topics:

Research

Maryland Today is produced by the Office of Strategic Communications for the University of Maryland community weekdays during the academic year, except for university holidays.