Judi Julmisse, a Black woman in Miami, was your typical Hinge user. She created a profile listing her relationship wants and her career. She answered prompts to show off her personality and uploaded recent photos of herself, all in the hopes of finding a good match.
But the men promoted to her on the app didn’t seem viable. Hinge showed Julmisse male users with grainy photos and personality-less profiles, she said. They often left off their dating intentions and didn’t fill out important prompts about their personalities.
“Why are these the guys who are swiping right on me?” she recalled thinking.
Then she saw a TikTok video that made sense of it all.
Odion Eigbobo, a Black woman in Arlington, Tex., said on TikTok she tested Hinge’s algorithm to find out whether changing her racial data would display better suitors. She set up two versions of her profile, identifying as Black on one and White on another, but using the same photo for both.
The White profile received higher-quality potential dates, she told viewers.
Julmisse followed her example — and also noticed what seemed like an uptick in suitable matches.
Throughout Eigbobo’s TikTok comments, Black women frustrated with dating apps recounted how they conducted similar tests — and pointed to the results as evidence of biased algorithms that block them from finding good suitors.
Matching puzzle
Hinge, a property of Match Group, allows users to filter matches by broad ethno-racial categories such as Black/African descent, White/Caucasian, Hispanic/Latino, Pacific Islander and more, spokeswoman Tamika Young said in an email. Users can also state that they are open to all races as potential matches.
It’s one of several choices that users can put into their profiles for free — along with sexual orientation, location, age — to reflect their personal preferences in a potential date. Subscribers or paid users have access to additional filters, such as height, political affiliation and education level. A user’s results are mostly based on these preferences as well as the profiles they either like or skip.
“We want to give our daters the ability to find others with similar values, cultural upbringings, and experiences that can enhance their dating journey,” Young wrote.
Black women and some researchers say this matching premise from Hinge and other apps can’t be the whole story if these women aren’t seeing good potential dates. The visibility of high-quality dates on White profiles reinforces their feeling that mainstream dating apps don’t serve Black women well — and that bias might hamper their pursuit of long-term relationships.
Hinge is one of the few apps that still lets users filter matches based on racial or ethnic categories. Many apps’ algorithms may have a bias that reflects real-life prejudice, these researchers say.
Hinge doesn’t reveal details of its algorithm because the information is considered proprietary. But algorithms tend to operate on the assumption that “like attracts like,” which can come down to users’ appearance, said Apryl Williams, assistant professor of communications and media at the University of Michigan and author of “Not My Type,” which explores racial bias in dating apps and the fetishization of minority users.
In other words, the apps might infer you want to talk to people of your own race based on assumed beliefs about coupling that are programmed into matching systems.
In her book, Williams reviewed the patent of Match Group — which owns other dating apps besides Hinge, including Tinder, OkCupid, Plenty of Fish and BLK — and concluded that the base algorithm could be the building block for all of its apps. This first layer may be inadvertently “siloing” users of color even for users who would otherwise be open to dating different races, everything else being equal, she wrote.
This silo effect affects Black women more because it can feed into stereotypes that they are less desirable on the dating market. These concerns came to light as early as 2014, when OkCupid released survey data showing that White, Asian and Hispanic male users rated Black women as less desirable compared with White, Asian and Hispanic women. Only Black men rated Black women in a favorable light.
This debate is heating up as dating apps are coming under fire more broadly for becoming increasingly pricey as well as implicitly promoting addictive behavior. As a case in point, a class-action lawsuit filed in February accused Hinge of creating “swipe addicts” and not living up to its slogan “designed to be deleted” — referring to its stated promise to get users out on dates that will eventually lead to serious relationships.
In response, Young declined to comment on the lawsuit but reiterated that the company is structured around the goal of getting people on dates.
It’s not just dating apps that struggle with bias. In an investigation by The Washington Post, AI-image generator Stable Diffusion XL demonstrated a bias toward young, thin, light-skinned and light-eyed women as being attractive. Robots programmed to scan people’s photos and use AI to mark which ones are criminals repeatedly selected Black men. And large language models such as Chat GPT4 labeled fictional speakers of African American Vernacular English as stupid, ignorant or lazy.
The switch
In her experiment, Eigbobo’s non-Black profile received twice as many likes — flags from other users signaling they want to talk to her — compared with her Black profile, she said. This initial set of likes included men who said they were ready for long-term relationships and noted their careers. Most of the users who liked her Black profile, by contrast, didn’t take the time to fill out the dating profile in earnest, she said.
Julmisse tried the same tack — in this case marking herself as White and Asian — and noticed better daters and more likes. The algorithm showed her profiles with clearly labeled dating intentions and career paths as well as better pictures.
Many commenters on Eigbobo’s 17 videos on the subject said they shared similar Hinge experiences. Among the nearly 650 comments on the first video, most who tried Eigbobo’s experiment on their own Hinge app had similar results, she says.
One commenter, who is nonbinary and spoke on the condition of anonymity to keep their dating life private from colleagues, had initially listed women and nonbinary as their preference on the Black profile — but were recommended men and couples. Once they switched to White but kept everything else the same, they got more matches that fit their marked preference.
Preference issues
Eigbobo’s video has also attracted critics. Some dismiss her claims of bias based on their own positive experiences with Hinge and other dating apps. Others claim that Eigbobo simply doesn’t match what other users are looking for, particularly if they don’t want to be matched with Black women.
Eigbobo’s rebuttal is that users are “trying to call her ugly, but it was never giving ugly,” she said.
The preference issue doesn’t explain the increased likes on her profile, she said. Users can see she’s a Black woman — even if their racial preference is marked otherwise. In those cases, in theory, they should bypass her profile.
Other users say experiences can vary by app. Jazmine Randle, a Bumble and Hinge user, received fewer matches on Hinge when she identified as White but saw more users that she found attractive. On Bumble, she did the same experiment of changing her race to White — and she received roughly the same number of matches but fewer successful conversations.
“I’m not actively looking and it’s frustrating. I can only imagine what it’s like [for others],” Eigbobo said. As for her two profiles, both have been on pause since June.