Unless You’re A White Guy, Facebook Thinks You’re Black

I am a white woman living in Minnesota, born and raised in neighboring South Dakota. My dad comes from a long line of Swedes who settled in the Midwest in the late 19th century. My mom’s heritage is a bit fuzzier, but as near as we can tell, it’s Northeast European/Scandinavian. I am white: I don’t tan; I burn. Most of the people in popular media look like me. My family gatherings are a sea of bland potatoes, cheese, and something called a “banana apple whipped cream salad.”

But Facebook thinks I’m Black. When I turn off my trusty AdBlocker, I get advertisements for Black dating websites and Black-owned businesses. A quick check of the ad preferences data that Facebook makes available to users reveals that I’ve landed in the category of “Multicultural Affinity: African-American (U.S.).” The U.S. is, of course, a redundancy, as “African-American” is already specific to this country. But this is just one of the ways in which Facebook’s algorithms are built with assumed whiteness in mind—there is no “multicultural affinity” for being white.

“Multicultural Affinity,” rolled out in Facebook’s world in 2015, is Facebook’s technically not categorization by race. There are three main options: Asian, Hispanic, and African-American. This is created so Facebook’s advertisers can target their ads by race as well as by age and gender. And, in 2016, Facebook got into hot water when it was discovered that advertisers were using this tool to purposefully exclude people of color from their ads. Though we can’t know for sure, because Facebook is notoriously obtuse, the social-media platform has allegedly eliminated this as an option, but the continued existence of such categories speaks volumes to Facebook’s approach to the world. Namely: Facebook’s default is to assume its users are white.

A brief poll of one of my college friend groups revealed this to be the case. In the small sample group of approximately ten white people, split fairly evenly along gender lines, revealed that our socially progressive—e.g., those engaged in anti-racist, intersectional politics—white women were more likely to be categorized as having a Multicultural Affinity for African-Americans. The white men in my group were not categorized with any racial affinities.

I repeated this experiment on my Twitter, where I have 7,800 followers. In a similar result, many white women, non-binary people, and transgender people found themselves categorized as “Multicultural Affinity: African-American” but again, Facebook had categorized very few cisgender white men this way. These are groups of marginalized people being assigned to other marginalized identities, however incorrect those identifications may be. It seems that Facebook simply looks for any deviation from White Maleness and proceeds from there.

The element that triggers this affinity categorizing is obscure, but it appears to be related to one’s “likes” and postings. The advertising group Nanigans writes that a person’s “likes” and interactions contribute to a whole picture of a person, as opposed to analysis of names or pictures. While I “like” a well-rounded amount of both white and black media, I also post frequently about white supremacy in the United States (see my previous work here at DAME). I also have a not-insignificant number of people of color as my friends, thanks to having traveled extensively and being friends with a lot of people from all over the world. But the fact that Facebook simply assumes that because I have friends who are people of color that I must also be a person of color is indicative of an entrenched white outlook in tech and media that defends and upholds white supremacy without even trying.

One of the major features of white supremacy and white racism is its insidious nature: It is the background radiation of so much that white people often act in oppressive ways simply because we assume whiteness to be the default. Any other race is considered ontologically and irrevocably Other.

Indeed, I was prompted to explore and research this piece after This Week In Blackness’ Elon James White tweeted about what Facebook’s advertising identity said about him. In the discussion that followed, a number of white women expressed that they had the same categorization of affinity for African-Americans, which was puzzling. White’s is a voice in Black media I’ve followed for awhile—I try to make sure I am listening to voices from backgrounds different from my own. Apparently, deliberately seeking out Black, Asian, or Latinx voices is enough for Facebook to assess my profile and categorize me as African-American, because their bar for media consumption and enjoyment is set at whiteness, while deviations from that norm are immediately distanced and placed into separate, distinct categories.

I believe white people who wish to be anti-racist have an obligation to listen to media before the majority-white mainstream. According to Facebook, this belief means I inherently have handed off my whiteness—that my whiteness is somehow not as clear or obvious because of what I read and listen to.

But as tech companies continue to face ongoing problems with race—facial recognition tech not being able to distinguish between different Asian people, the Apple Watch’s sensors not working properly on brown skin, even basic elements like photography not having the technology to properly light and photograph brown  skin—Facebook’s clumsy categorization of the Other is emblematic of tech’s larger problem with whiteness.

Facebook’s own user base is largely people of color—10 percent of Facebook’s users are based in India, the second largest user of Facebook behind the U.S., at 12 percent. What’s more is that Black people in the U.S. are more likely than white people to be on Facebook, with Hispanic people outpacing them by about 3 percent. But Facebook still caters largely to white people with the assumption that white people are the default. This may simply boil down to Facebook’s own internal demographics: as of 2015, the majority of Facebook’s staff were white and male.

This is the hidden power of whiteness—Facebook doesn’t necessarily see any problem with its algorithms because categorization of non-white people as Other is embedded into the data. Engagement with non-white media is evidently so uncommon amongst white people that your behavior will get you shifted into an entirely different race if you choose to widen your media diet. Go beyond whiteness and you are immediately Other. And that contributes to the continued assumption that Black people can be expected to consume white media and relate to it, but that media from people of color is only for people of color.

And because whiteness dominates mainstream media, this further stigmatization of media helmed and developed by people of color only serves to contribute further to the divide. White media is “general.” POC media is “specialty.” And that, ultimately, is the problem. Until white people as a whole are able to understand and reject this idea that media created by POC is some kind of specialty thing, we’ll never be able to move past the dominant narratives of whiteness that plague national news: everything from the Austin Bomber being characterized as a sympathetic loner to Donald Trump getting the benefit of the doubt when he manages to read a teleprompter without screwing up, or gets through a meeting without the prime minister of Japan without devolving into commenting on racist stereotypes. None of these problems will change until we start incorporating perspectives other than whiteness into our media.

But this is more than just a problem of tech or of media. This reaches beyond the fact that, as the Avenue Q anthem goes, everyone’s a little bit racist. This is fundamental to the ways in which we move through the world of media and technology: the presumption of whiteness until blackness is proven, rather than assuming a natural diversity of perspectives and viewpoints from the start. This is a fundamental problem of whiteness propagated through media. And noticing it is the first and hardest step toward fixing it.