TikTok beauty filters can be super realistic—unless you're a person of color

Stay updated with the latest in Tech, Science, Culture, Entertainment, and more by following our Telegram channel here.

When I first came across the Glow Look filter on TikTok, I dragged my self-absorbed ass into the video maker to try it on. I was expecting to be wowed by my own hotness, just like all the other girls on my For You Page.

Instead, I tapped the filter on, and immediately guffawed. It looked ridiculous, to say the least. The enlarged, bright blue eyes, flushed cheeks, and noticeably thinner nose looked extremely out of place on my Vietnamese face.

I returned to scrolling and quickly noticed two types of videos using Glow Look. The first group was mostly white women who were indeed wowed by their own hotness. The second was mostly women of color, either disappointed that the popular filter didn't really work on their faces, or protesting the Eurocentric beauty standards that the filter upholds.

"Honestly, my first reaction was like, 'Oh, great, another one of those beauty filters that changes our features to make us cater to the European so-called beauty standards,'" said Himani Jadeja, a TikTok creator whose content focuses on Desi lifestyle and culture. "It's so damaging to those who don't conform to those beauty standards. There's other ethnicities and other cultures who have their own form of beauty. And we just, I guess, don't match up to those standards."

"It's so damaging to those who don't conform to those beauty standards."

Jadeja isn't the only TikTok user who took issue with the filter. Videos from white creators praising the filter's effectiveness are peppered with comments from people of color pointing out that the filter is only impressive to white people because it was made for them.

It's an extremely othering problem — especially as face filters on TikTok, Snapchat, and Instagram become more realistic. When it's on the faces they're designed for, like in this TikTok by user Whitney Hanson (screenshotted below), the filters' dramatic changes are much less noticeable. And that makes sense when placing Eurocentric features on a face that already conforms to European beauty standards.

But when the same filters are applied to non-European features, the previously hyper-realistic effect becomes comical. It's obvious the "makeup" look is digitally created, and altered facial features are clearly out of place. When these filters blow up, entire populations are unable to participate.

Glow Look has been used in more than 3 million videos, but the phenomenon isn't isolated to TikTok. Similar facial beauty filters blow up on Instagram – like ATTRACTION, which has been used in more than 143,000 Reels. They too create what appear to be European-inspired features, often with lighter-colored eyes, lighter skin, and thinner noses. As far back as 2016 users noticed a skin-lightening effect in several filters on Snapchat.

"There's a lot of surface level interaction with this [filter], because [most people] are using it as a 'catfish challenge,'" said TikTok user Moe Khine, who identifies as Myanmarese. "Meanwhile, for people who might have a different face shape, it's not really a challenge for us. I looked like a spotted white walker."

So who's making these filters?

In a LinkedIn blog post LinkedIn blog post from 2018, Lu Wang, TikTok's head of AR effect design, broke down how the user design team creates filters for the app.

"After the developers have completed the background coding and attached it to the design, we will then use an in-house app to test and ensure that the filter's functionality is optimized," Wang said. "After this testing, the filter is launched to millions of users to add to their videos."

In the blog post, Wang doesn't reveal how the facial filters are trained. Most beauty filters use something called deep learning, in which a computer is trained to recognize facial features from photographs of real faces.

If TikTok's filters use a similar learning process, it is the company's responsibility to ensure the computer is trained on a diverse set of faces.

We asked TikTok for further comment on how the company develops its filters, and will update this story if it responds.

It's not just about feeling left out

While beauty filters are eye-catching and impressive, their constant presence can be mentally taxing for both active users and casual viewers.

"It depends on what people [may] have predispositions towards, and what their perceptions around [beauty standards] are to begin with," said Allycen Kurup, a graduate student in Purdue's clinical psychology program whose work focuses on adolescent digital communication and its effects on mental health. "But I definitely think that these filters continue to contribute to our overall culture of upholding and expecting certain beauty standards of people, and it puts it into this virtual reality that lets you pretend that you are meeting these beauty standards."

Using and seeing a filter that's meant to make you look "prettier" has the potential to be harmful to any user's self image. But for people of color, when something that's supposed to make us look "better" can't even work on our faces, it suggests that there's something inherently wrong with our features.

"It's just perpetuating this very ideal of ideals. You have to be a white woman. You have to have darker skin almost, but in the, bronze-y, 'white woman with a tan' way rather than like, actually working for people with different skin tones," said Khine.

In reality, of course, this isn't a facial features problem. It's a tech design and societal standards problem. The people behind TikTok (or Instagram or Snapchat) filters have a responsibility to create them for every audience.

"All of us are beautiful, too," said Khine. "[TikTok] needs to not only expand its way of thinking about beauty, but also about how they design these filters in the first place."

Follow Mashable SEA on Facebook, Twitter, Instagram, and YouTube.