The algorithms used in dating apps steer which user is shown to another. The programs study the behavior of the individual and other users on the app. But the results aren’t always fair.
Category
🗞
NewsTranscript
00:00Technology is changing a lot our love life because it has introduced new quantification practices
00:07that we didn't have before. So before we had maybe a matrimonial agency or friends
00:14introducing us to their friends, we went to bars. Today we have learned to swipe, to give stars or
00:22ratings and quantify in a very specific way who we are and what we like so a computer program
00:30can understand it. That's because algorithms dictate which user is suggested to another.
00:37The end goal is a match and hopefully a date. To find suitable candidates the algorithms learn from
00:45user behavior. Let's assume Hannah is looking for a partner. If she gives Akash a like the algorithm
00:54scans other people he's liked and who have liked him. The assumption men who like these women might
01:00also like Hannah so they are suggested to her too. Much like the recommendations you'll find
01:06on Google or Amazon, dating app algorithms base their suggestions on the decisions previous users
01:12have made. For example let's say many people give Akash a like but only a few like Duke. If I give
01:19Akash a like the algorithm will assume that I, like most of the people before me, will reject Duke.
01:25So Duke doesn't even appear as a suggestion in my feed and doesn't get any likes.
01:33The more popular a profile the more it is suggested to others. The unpopular ones remain unseen.
01:42In a survey among users the company behind dating app OkCupid identified a clear racial bias
01:49particularly against black women and men and Asian men. These groups were often rated lower than
01:55others. The system is designed to favor the preferences of the majority. Let's say on Tinder
02:03we have a majority of people that are white skinned. They're gonna have a specific preference
02:09patterns overall in average and that's also source of discrimination because Tinder is
02:15calculating the average of preferences. If the majority of people there is white people they're
02:21gonna create discrimination patterns towards minorities. If these white people are heterosexual
02:30then LGBTQ plus community is going to be discriminated. So it reduces diversity.
02:37Racial prejudice is embedded into our feeds and it can be reinforced consciously or unconsciously
02:44just by swiping. If you want to break through a bad ranking finding a partner can become a question
02:50of money. You also have to pay sometimes for features and one typical feature of this system
02:59is the boost in Tinder for example. If you pay they say they will put you at the top of the results.