Science fiction: what will be the near future of online dating

Фото: Depositphotos

The correct analysis of likes in social networks is able to tell about a person and his potential partner much more than the most honest questionnaire. Very exciting article from Gizmodo, which tells not even about the future, but about the present online dating services. Material translated by project Newly.

When I enter my twitter nickname in the dating application Loveflutter, it rewards me with a 28-point stat. According to the analysis, I am a type A - neurotic, anxiously obsessed with sex (99th percentile - a measure in which the percentage of total scores is equal to or less than this measure - approx. Newly). In the sidebar, where a photo of my profile is displayed, there is a subsection “Dating Tips”: “Try not to be negatively tuned. Get down to business right away, do not waste their time. Their patience may run out if you act too slowly. ” I'm still the one catch.

Loveflutterdeveloped in the UK based on data from Twitter and does not oblige me to fill out a personal questionnaire or to figure out what to write in the “About Me” section (the application does it for me - just in 140 characters). The application works in conjunction with a company analyzing language materials. Receptiviti.ai. They calculate compatibility between users through databases, as well as matches in the news feed. Twitter. Is this a good matchmaking method or is it just a trick? I think my position - anxious neurotic - is obvious.

Dating apps promise to instantly bring us together with those with whom we are destined to be together. Sometimes it works, sometimes it doesn't. And since the algorithms of machine learning are now more accurate than ever, the dating companies will be able to find out who we are and with whom we go on dates. Familiar dating sites will soon be in the past. The coming mercilessly, but halfway already passed.

Фото: Depositphotos

Evaluation of "personality"

Rating companies are divided into two categories. Type Services eHarmony, Match and OkCupid asking users to write voluminous essays, as well as fill out questionnaires. Based on the data entered, the site selects a pair for the user. Such questionnaires are informative, but their completion takes a lot of time. In addition, there is a desire to embellish information about yourself by answering the questions "Do you play sports?" Or "Can you call yourself a slob?". On the other hand, companies like Tinder, Bumble and Hinge miss the moment with a long filling of questionnaires. Instead, the user is asked only to give a link to the account on the social network. tinder fills in a new user profile automatically using music from Spotifyphoto of Instagram, friends and likes from Facebook. the main task tinder - do not analyze compatibility, but constantly attract new users who choose their own pair, looking through brief information.

It is true that we reveal much more than we realize, in our tweets, posts like Facebook, photos in Instagram and marks on foursquare. We give dating apps access to this data and more. When a journalist from The Guardian requested from tinder all information about itself, the company sent a report on 800 pages. Sounds scary? Maybe. But when I worked with databases in OkCupid, such huge amounts of information delighted me.

Фото: Depositphotos

In the future, applications like tinderprobably will be able to tell in detail about the properties of the person and the current lifestyle of users through the analysis of activity in social networks. They will provide more information than the questionnaire eHarmony ever be able to accommodate. Researchers now believe that they can judge the state of a person by his behavior online: about neuroticity - according to Chekin foursquare, about the presence or absence of depression - on tweets and the choice of filters in Instagramand likes Facebook - about erudition, contentment of life and propensity to take drugs.

In addition, our online behavior is often illogical. In 2013, the University of Cambridge analyzed the relationship between the user's personal characteristics and his likes on Facebook. The results showed that intellectuals preferred such pages as “Science” and “Colbert's Report” (a satirical TV program on Comedy Central — approx. Newly), which is not surprising. But beyond that, they liked Thunderstorms and Spiral-Shaped French Fries. It may sound strange, but does the logic have any significance if the algorithm for determining personal qualities is used only to find the right partner?

Watchdog oversight

Many traits are almost impossible to recognize. As a rule, we are not so diligently monitoring our own profile. Facebook, like a page on a dating site. Perhaps, therefore, the data Facebook will be more accurate than the results of a voluntary survey.

"On dating sites, people would rather show themselves in a more favorable light, that is, unrealistic," says Chris Danforth, a professor of mathematical, natural and technical sciences at the University of Vermont, who studies the patterns between posts in Instagram Twitter and depression. “But every time we receive the results of such research, we understand that we are talking about ourselves much more than we realize. Maybe we are silent about our preferences, filling out questionnaires, but our actions speak volumes. By likes Facebook it is possible to determine more precisely whether people get along with each other than with questionnaires. ”

Also, social networking data can be used to motivate users to stay honest when creating a profile. “I think it would be interesting if OkCupid checked users when filling out personal information, reflects Jen Golbeck, a researcher at the University of Maryland, who studies the intersection of social networks and user information. “He could have said something like:“ We analyzed your likes and came to the conclusion that you may be smoking. ” Are you sure you want to choose this answer option? "". Most downloaded dating sites could even warn the user about a possible distortion of the facts.

Based on the action data, companies could flag suspicious profiles with red flags and thereby warn people of possible danger. After the rally of nationalists in Charlottesville in August of this year, some online dating services asked users to provide the addresses of the neo-Nazi pages and blocked them. But in the future, sites may be able to independently identify sexists / racists / homophobes and so on by their activity in social networks and block pages of suspicious individuals. (It is possible that such functionality would solve the problem of harassment on dating sites.)

Фото: Depositphotos

But it is also possible to block users allegedly unsuitable for relationships. The eHarmony service, for example, rejects candidates who have been married more than four times, as well as those whose questionnaires indicate a tendency to take an example (discrimination against people with chronic diseases and disabilities - approx. Newly) or depressed. In the dark future, dating sites may learn to identify users who suffer from anxiety by likes or posts.

According to our behavior in social networks, it will not be difficult for the algorithm to find out real answers about which we can lie. For example, in OkCupid There is a question of how much you train. But MeetMeOutside, a dating site for sports people, asks users to connect sensors and prove physical activity by counting steps. It's harder to embellish. Or, instead of asking the user what they like more - watching TV shows at home or having fun at a nightclub, the service can determine this using GPS and foursquare, then to find him a suitable pair.

Worship algorithm

Computers have access to a huge array of data and impressive computing power, which allows you to identify patterns that a person is not able to notice. “When you view photos of people on social networks, you are able to analyze the action and behavior of a particular person at a particular point in time,” Danforth said. “But the algorithm will be able to study the differences between its behavior and the behavior of another million people. When viewing a tape, you on an instinctive level feel something that is impossible to grasp, an invisible dimension, non-linear combinations that are difficult to explain. ”

Analytical algorithms for data services are constantly being improved in the study of the personal characteristics of users and their preferences, without asking unnecessary questions. Some applications already use these patterns, analyzing who you swipe to the right, and whom you swipe to the right, just like Netflix recommends you movies based on the ones you liked before.

Фото: Depositphotos

“Instead of asking questions about individuals for users, we work exclusively on analyzing their behavior on a dating site,” said Gawin Potter, the company's founder. RecSyc, which algorithms are used by dozens of dating applications. - We do not ask „What people do you like? From 50 to 60 years? “, But we are watching who the user is looking at. If these are 25-year-old blondes, our system starts recommending 25-year-old blondes to him. ” Data OkCupid show that heterosexual men tend to write to women significantly younger than they indicated in their questionnaire. Therefore, recommendations based on behavior analysis are much more accurate than user-specified settings.

Algorithms that study user behavior can also recognize subtle, inconspicuous patterns, based on which you can determine the type that you find attractive. At least, some application developers say so.

“If you look at individually selected recommendations, you see everyone who corresponds to a certain type: brunettes, blondes, people of a certain age. Some women in Houston prefer men with a beard or mustache. We found users in China who prefer very, uh, modest people, ”Potter informs me, apparently implying a stereotype I don’t guess. “None of the questionnaires I know fix this.”

Naturally, the computer-generated attractive type of appearance may not be pleasant to us. When I asked Justin Long, the founder Bernie.ai, what patterns were discovered by their neural network, he didn’t reveal the cards: “We received some alarming results that I don’t want to share. They are pretty offensive. ” I guess they were racist: statistics OkCupid shows that even if people they saythat they do not care about skin color, they choose a partner among a certain race.

“Personally, I think that my style of svaypas and the people I choose as the most suitable cannot indicate the presence of hidden personal qualities,” said Camilla Cobb of the University of Washington. She explores dating services in terms of technology and privacy policy. - We simply use applications to search for people that interest us. I do not think that the data companies are studying the data in order to harm my reputation, they use analysis to find matches. But if I do not want to know about my hidden prejudices, then I will not use this kind of application. ”

Data thirst

Even if dating services do not use our data to damage reputation, they can be used for profit. It’s hard to imagine what data they can provide to advertisers, especially if this information that we don’t know about ourselves ... I don’t smoke, but if I only have the guys with cigarettes in the photo, then most likely I think that a man with a cigarette looks cooler. The advertiser can find out which products we subconsciously consider to be sexual and will show targeted ads.

At the moment, compatibility analysis algorithms tend to encourage users to swipe right. As applications are improved by analyzing the user's identity and preferences, in a short time the user will no longer need to post, like, and exchange messages on their own. Thus, the Canadian engineer Justin Long thought in the process of creating Bernie.ai - “personal mate assistant”. Frustrated by the amount of time spent on svaypas and the exchange of messages that precede rare real dates, Long created a bot that does a boring job for him. application Bernie asks the user to link to your account in Tinder, and then tracks the swipe, studying the individual tastes of users. Then Bernie starts flipping tinder instead of user. If a bot notices mutual sympathy, he begins a conversation with the phrase “Do you like avocados?”.

Finally tinder forced Long to stop developing the bot, but the programmer still believes that personal assistants like Bernie - the future of dating services. Instead of spending time scrolling and messaging, we will give virtual assistants access to calendars and location and let them negotiate. “My Bernie will talk to your Bernie,” Long says, and the bots will automatically make a date. When the methods are so good, we trust their decisions, but perhaps we allow unnecessary control of privacy.

You're on your own

As the algorithms improve, they will need data not only from those users whose avatar we rate, but also from those to whom we feel attracted in real life. None of the dating apps (known to me) request data from users about the results of these dates. When I asked this question to the main developer OkCupid Tom Jacques (to his ex-boss), he evasively replied: “This is a delicate subject. There are things that people are not willing to freely share, and we can only track interactions on the site. At some point, they transfer their communication to real life, and not everyone agrees to tell us about their experiences (successful or not). ” However, we provide applications with enough information to determine how these dates went. Services can use our GPS coordinates to observe with whom we went on a date, how long it lasted and whether there will be a second one. application Once even allowed users to monitor their heart rate on a date through Fitbitsto find out how excited the meeting atmosphere is. (Although Rosalind Picard, an expert at the Massachusetts University of Technology in recognizing emotions through biosensors, told us that changes in pulse rates reflect body movements rather than the slightest emotional fluctuations.)

Today, dating applications do not receive (openly) our digital data to the maximum extent possible. Perhaps, the e-mailing companies believe that such monitoring will seem too creepy to users or that they will be outraged by the excessive awareness of service employees about the facts of their personal lives. But if data mining would be the prevention of failed dates, wouldn't it be worth it?

I still have doubts, but the more I like the idea of ​​a hyperintellectual insightful algorithm for dating, the more I think about deleting my account on Loveflutter.

Read also on ForumDaily:

Why creators of popular technologies refuse to use them

Facebook happiness: how social networks reflect reality

How Facebook reveals family secrets and what the social network knows about us

Big brother is here and his name is Facebook

Do you want more important and interesting news about life in the USA and immigration to America? — support us donate! Also subscribe to our page Facebook. Select the “Priority in display” option and read us first. Also, don't forget to subscribe to our РєР ° РЅР ° Р »РІ Telegram  and Instagram- there is a lot of interesting things there. And join thousands of readers ForumDaily New York — there you will find a lot of interesting and positive information about life in the metropolis.