Connect with us

Technology

When Love and the Algorithm Don’t Mix

Published

on

/ 5605 Views

When I met my husband, who happens to be white, he told me that he was always seeing women with blonde hair on Tinder and he’s not really into blondes. No matter how many times he had swiped left on blondes, the algorithms were always recommending them to him, presumably because pop culture dictates that white men prefer blondes. Luckily for us, the algorithms’ tendency to stack blonde women in his swipe deck worked out in our favor because I’m a black woman who, at the time, had blonde hair. 

In nearly 10 years of swiping through profiles on Tinder, Bumble, Hinge, and OkCupid, I learned that dating apps can provide pathways for finding friendship, adventure, romance, and sometimes, love. But there was one aspect of dating app culture that I couldn’t ignore because it was often the first thing matches wanted to talk about: race. People wanted to know where my skin tone came from. They asked if I was mixed. They wanted to tell me that they’d never been with a black woman before.

Online dating as a black woman is a unique experience. Black women are often over-targeted by those seeking exoticism, yet we are told that we are the least desired dating demographic. I wondered how dating apps’ algorithms reconciled my blackness: was I being placed in match decks because of my blackness or despite it? It also sparked another question: Why do so many daters still wind up single after spending countless hours on dating apps? Maybe the algorithms just aren’t into them, especially if they are people of color or someone who dates folks of all ethnic and racial backgrounds. Or perhaps users’ preferences keep them from matching with others from beautifully diverse backgrounds. In fact, the cofounder of OkCupid , Christian Ruddersuggested in his book, Dataclysm, that race has no bearing on compatibility, rather, opinions about race have the ability to make or break matches. The answer, as our research shows, is all of the above.

Read More: How Rizz Assistants and AI Matchmakers Are Transforming Dating

As a sociologist who studies race, gender, Technology, and popular culture, I have interviewed over 100 people about their experiences with online dating, with approximately 75% identifying as people of color. Many shared that swiping and chatting left them unfulfilled, lonely, and often, healing from trauma—racial trauma. I, alongside fellow researchers from the University of Michigan’s Departments of Psychology and Communication, conducted an additional study that, among other things, asks why online dating as a person of color is so fraught. Preliminary findings suggest that people suspect algorithms used in dating apps are a major reason why. One interviewee, Grace (her name has been changed to maintain privacy) shared, “When I first jumped on there, they did try to connect [me with] people who look like me.” Grace was onto something. 

The idea that physical similarity is necessary for an ideal match is rooted in centuries of anti-interracial mingling ideology that can be traced as far back as 1661 in Virginia. Laws banning interracial marriage carried a fine of 10,000 pounds of tobacco. White women who had mixed race children had to either pay a fine or submit to indentured servitude for five years while their offspring were committed to 30 years. These heavy penalties demonstrate how deeply invested early Americans were in maintaining racial purity; a legal enforcement that prevailed until 1967, when the U.S. Supreme Court ruled that such laws violated the 14th Amendment rights to due process and equal protection under the law. But History has demonstrated numerous times, the ending of a legal practice does not swiftly coincide with social acceptance of that practice.

Politely coded as personal preference for some, ideas about intimate racial mixing are one of our last hushed taboos. Most don’t mind mixed-race coupling, as long as it does not occur in their own families. Though, as a collective, we may appear to move away from the reality of racial injustice associated with prohibiting interracial marriage, the discourse of racial purity is an ever-present tenor in American Politics and in our technologies. 

Match Group, the parent company to Tinder, OkCupid, and Hinge has filed a series of patents suggesting that the relevance algorithms powering their dating apps select on hair color, eye color, and ethnicity. A deep dive into Match Group’s patent, which outlines the mechanics of their matching and sorting systems, supports the belief that many people have: The algorithms try to connect daters with people who look like them. The patent states quite plainly that “people having similar and/or compatible character traits and values should be matched.” Match Group’s patent also indicates that a “relevance algorithm” may use signals to evaluate similarity between daters. These signals could also include characteristics such as “heigh, weight, age, location, income, and ethnicity.” Where did Match Group get the idea that shared physical traits equate to similar values and character traits? From us.

The dating industry is an extension of traditional matchmaking. Matchmakers are skilled at reading the cultures in which they work. Online dating companies use algorithms to make predictions about individuals supported by an adept knowledge of global cultural flows and trending idiosyncrasies. They treat physical traits as proxies for cultural ideology and symbolism because we do. The apps use algorithms that amplify preferences for particular body types, racial presentations, or height because we speak this way on the internet and with our closest friends.

Most people are not public about their racial preferences in partners. In fact, we have a secret language for talking about racialized beliefs about attraction and intimacy: we talk about “type.” Everyone has a type: a mix of physical, emotional, and political characteristics that generate some resonance for each individual. We may think of some of these aspects as a subjective matter of personal preference. But the truth is, perceptions of an ideal type are largely shaped by cultural signifiers and beliefs about race do a lot of work to hone those preferences. Our parents, religious beliefs, schools, and socioeconomic standing all play a part in shaping who we find attractive and desirable. For example, evidence suggests that conservatives prefer thin white blondes, perhaps because this type of woman signals a performance of class and race  that is a useful currency in some conservative circles.

This outdated superficial matching, based on physical similarity, may work for some, but it misses the mark for many daters who are seeking to connect with others around shared values such as approaches to Health and safety during a pandemic or alignment on climate change . Why? Perhaps the online dating industry has read the culture so well that they know our secret. We purport to be liberally minded daters who prioritize our values above all else. Yet, the hushed taboo of sexual racism, defined as personal racialized reasoning in sexual, intimate, and/or romantic partner choice or interest, connotes a set of beliefs, practices, and behaviors that provide commentary on what is considered socially acceptable desirability. Sexual racism presents a barrier to meaningful connections when we can’t see past stereotypes about groups of people.

If we think of the dating industry as a mirror of social truth, quietly reflecting sexual racism, online dating companies’ outdated approach to a socially stratified society is unsurprising. The ideas which shape and drive online dating culture, and the tech industry at large, come from a society that routinely fails to deal with social inequity at both systemic and individual levels.

We construct belief systems about values, norms, and standards which pervade algorithmic design but technologists and ethicists have repeatedly demonstrated how the tech industry struggles to untangle design and innovation from America’s racially unjust reality: large language models fail to address racism in their data, recidivism algorithms incorrectly identify black defendants as future criminals, and algorithms used in Healthcare can lead physicians to misdiagnose patients because their systems rely on dated racially biased data. Where dating apps are concerned, industries and governments have historically invested in promoting partnerships of racial sameness—an ethos that in some circles, still quietly lingers on. 

If dating apps continue to follow the tech industry’s trend of building algorithms that accept and amplify outdated forms of social injustice, they will fail their burgeoning and diverse user base. Not because these users are more diverse and cosmopolitan than those before, but because the online dating industry shapes popular cultural ethos about dating, hookups, and relationships. The behaviors that are established in the privacy of our devices seep into real world discourse on acceptable first date behavior, income thresholds, gender roles, height preferences, and so on.

The problem is that some of us don’t desire superficial similarity. Some of us do strive for deep bonds that transcend social ideals about race, body type, and upbringing. The apps, the algorithms, and the culture they reflect make it incredibly difficult to be an outlier in today’s dating culture. Dating companies should design for this growing margin of users in service of a better experience for their entire user base, and frankly, for the sake of our culture.

Currently, dating apps’ algorithms are working for the status quo. These racially biased algorithms work to successfully match the mainstream user with normative tastes. When we operate in the unexamined frame of normative desirability, our swiping behavior sends feedback to the apps, suggesting that we like this limited frame they’ve established. More important, when we refuse to examine our own prejudices, we may miss the perfect match. If we want to change racially biased algorithms in dating apps, we must first change our own racially coded evaluation systems.

Trending