Connect with us

Technology

Smarter AI, DNA tests, and video

Published

on

In our Love App-tually series, Mashable shines a light into the foggy world of online dating. It is cuffing season after all.


Digital courtship exploded with the rise of the smartphone. 

Match and eHarmony laid the online groundwork decades ago, but momentum built after the first iPhone was released in 2007: Grindr was founded two years later, Tinder in 2012, and Bumble in 2014. 

These apps, bolstered by location-tracking, swiping, and almighty algorithms, brought the masses to online dating. But as we look to the future, online dating companies have a new problem to tackle.

“I feel like we’ve solved the volume problem,” said Hesam Hosseini, CEO of online dating behemoth Match (Tinder, Hinge, OkCupid, and Plenty of Fish are all owned by Match). “How do you solve the quality part of that volume equation? If we’re going on a lot of dates, great, but are we really on a better path to finding a partner?”

“How do you predict chemistry? I think it’s hard but I think we can get there.”

App innovations and society’s increasing comfort level with online dating have built large pools of potential dates. But a fix to the quality issue remains to be seen: Will we be going on VR dates in 2030? Will we have digital butlers speak to our matches for us, weeding people out … or Siri-like matchmakers talking us through our options? In 2040, when 70 percent of couples are expected to meet online, will our phones show us, in augmented reality, how compatible we are with passersby? 

Hosseini and other execs I spoke to about the future of online dating don’t have imaginations as wild as Black Mirror fans would like. But their insights about what’s coming down the pipe — from better machine learning to video — hint at what daters have in store. One thing is for sure: Online dating isn’t going away any time soon. If anything, it’s likely to become further integrated into even more people’s lives.  

Machine-learning love

Online daters are exhausted, disappointed, and feeling jaded, now more than ever before, said Dawoon Kang, co-founder and co-CEO of Coffee Meets Bagel, the app made famous by rejecting Mark Cuban’s $30 million purchase offer. Machine-learning and AI may be able to help.

“People, a lot of times, don’t know what they want. They may say they want certain things, but they don’t actually want that,” Kang said. Better machine-learning could tailor your matches to your actions, rather than your stated desires. 

To better select matches, smarter AI could, for example, take into account how much time you spend in the app, as well as the profiles you’re looking at and for how long, the content of your chats, how you tend to swipe, how long it takes you to respond to certain messages, and whether you initiate chats. If you let this future online dating AI have free-range on your phone, it might even analyze whether you pay your bills on time, what websites you visit, the news you read, which shows you binge-watch, how much you exercise (if you have a connected Fitbit or Apple Watch), and your camera roll.

AI with improved photo detection could one day pinpoint interests that may not be outlined in your profile or use facial recognition to find your “type.” Data collected from everyone else’s matches — both successful and unsuccessful — could help improve your matches, too.

“I didn’t set out to build AI necessarily. I set out to build something like a human.”

“How do you predict chemistry? I think it’s hard, but I think we can get there,” Hosseini said.

Dating apps seem to already be paving the way.

Loveflutter, a UK dating app, has AI that matches people based on personality traits it decodes from their tweets. It also plans to use AI to coach users through meeting offline after analyzing their chats. Going further into the coaching arena, Match launched Lara last year. The digital personal assistant is activated by Google Home and suggests a daily match as well as dating tips and activities. Then there’s Badoo’s creepy Lookalike feature, which uses facial recognition to match you with people who look like your favorite celebrity.

Beyond all that is AIMM, a voice-activated dating app which launched last year and has 1,000 users in Denver (it’s planning to expand throughout the U.S. in coming months). An AI matchmaker, which sounds like Siri, asks you questions for a week before sending you matches. Along with those suggestions come personalized photo tours and audio snippets of your match describing their perfect date or telling an embarrassing story from childhood. There’s no tapping or swiping. Once both you and your match have agreed to chat, AIMM will set up a phone call, and you decide from there if you want to meet offline. 

AIMM will throw in a joke now and then as it talks to you, too, said Kevin Teman, AIMM’s creator. It can also pick up on your values through subtle conversations. For example, if someone talks a lot about money, AIMM could infer that money is important to them. 

“I didn’t set out to build AI necessarily. I set out to build something like a human,” Teman said, adding that AIMM remembers your previous answers and the tone and questions you warm up to. 

For Teman, there’s no end in sight to how much AIMM, and other AI, can learn. That may sound promising, but AIMM’s promotional video is pretty awkward.  

Jean Meyer, the founder and CEO of European dating app Once, doesn’t think the dating industry will crack the AI code. Rather, he sees a researcher from a deep-learning lab making strides here — and then struggling to get a dating app to buy in.

“The dating industry has zero incentive to create this perfect matching algorithm,” Meyer said. If the AI is too good, potential daters will leave platforms, depriving apps of the customers they need to make money. 

“The optimum for a dating service is to show you profiles of people that might be good enough, but not perfection. People who you go on dates with, and then after two, three dates you realize it won’t work out and you go back to the app,” Meyer said. 

Hosseini dismisses that line of thinking, noting that if you produce a lot of dates, but not a lot of relationships, that hurts your bottom line. App use is often driven by happy couples singing the gospel, after all. 

Gene matchmaking

People are already comfortable sending their spit to a lab to decipher their family history; more than 26 million people have done so, according to an MIT study.

But would you swab your mouth for a date? Some in the online dating field are betting on it. 

Pheramor, DNA Romance, and Instant Chemistry all analyze users’ DNA to make matches. (Pheramor also trains AI to mine your social media profiles to understand your personality.) The gene-matching evangelists propose that certain genes connected to your immune system, known as the major histocompatibility complex (MHC), govern who you’re attracted to. As with the cliché, the thinking goes that opposites attract; those with dissimilar MHC are expected to like each other. 

A few studies somewhat back this up — one involves mice and another sweaty T-shirts — but there are plenty of scientists who call this bunk. 

Still, eHarmony expects a flurry of lab-made romances by 2025, according to a 2018 report about the future of dating. 

“By analysing MHC gene codes, online dating platforms could close the gap between predicting whether their subscribers who initially are attracted [to] each other online, will still be attracted to each other, once they get offline,” eHarmony wrote.

Cracking video

OK, now for something a little less sci-fi. Video will play a larger role in dating apps, execs said, but how, exactly, is still unclear. 

Last year, Tinder introduced Loops, two-second videos used in place of profile pictures that look like Boomerangs, and Badoo added a live-video chat feature you can use to talk to potential dates. Coffee Meets Bagel has short-form video, prompting users to post their answers to daily questions on their profile. While lots of people like to watch the videos, most don’t want to create their own, Kang, of Coffee Meets Bagel, said. 

What Tinder Loops look like.

What Tinder Loops look like.

“It feels a little bit scary to do video chat with someone you haven’t met, and I think that fear is stronger among women,” Kang said. She’s confident that eventually video integration within dating apps will become normalized, but it’ll take time.

“We just have to overcome this mental block that it’s weird,” she said. 

VR and AR: The jury’s still out

Virtual reality in dating apps will also take cultural change, Kang said, but she thinks it’s the future — once everyone feels comfortable strapping on a VR headset. 

She sees scenarios in which daters meet in virtual reality and then play games together using augmented reality, or travel to one of the pair’s favorite destinations (all the while still sitting on their couches).  

“There’s no doubt in my mind that we’ll get there,” she said.

Meyer, from Once, likes the idea of VR, but thinks it’s too complicated. 

“Why is Tinder so successful? Because you can sit on your sofa, not giving a shit, take out your cellphone and start swiping people,” he said. “But if you have to set up your whole VR helmet and plug it in just to go on a date, how many people will have this incentive?”

Instead, he imagines that in 20 years time we may have AR glasses that show you who is single and a compatibility score above their heads. Given the fate of Google Glass, AR glasses seem unlikely, but using your phone to screen passersby in AR is a compelling thought. You could point your phone at someone in the distance and their dating profile would pop up, similar to Google Lens‘ shopping or search features. The initial ingredients are already stewing: Facial recognition is on the rise, and Google’s been flexing its AR muscles

But Kang also sees a very low-tech approach budding: dating apps hosting meetups for users. Last year, Coffee Meets Bagel held a 5K Singles Run in Hong Kong for 1500 people, a three-day summer camp that brought 150 singles to the backwoods of Pennsylvania, and monthly concerts themed around relationship topics in New York. Coffee Meets Bagel doesn’t have plans for 2019 events yet, but Kang sees potential. 

Flirty chatbots

As for digital butlers talking to our matches’ chatbots — which seems like the ultimate loss of humanity in dating, just after pressing a button to indicate you want sex — dating app execs aren’t convinced. 

Chatbots work for tracking packages and ordering pizza, not getting to know someone who could be your future life partner, Hosseini said. 

Bernie.ai, a chatbot with deep-learning AI and facial recognition that could sift through Tinder matches and strike up introductory text chats, launched in 2016. The bot could learn your preferences and talk like you, but it was short-lived. Tinder, unhappy with the AI, according to Motherboard, banned it in 2017 — after it had performed about 9 million actions, such as swiping, and made around 100,000 matches.

Bernie.ai’s creator, Justin Long from Vancouver, may have thought it was a success, but Hosseini said chatting with a potential date online is part of the courtship process. 

“Personally speaking, if that’s the future we’re looking at, it’s a scary place,” Hosseini said.

What everyone wants

That said, the dream innovation for dating site execs, one not limited by tech constraints or cultural adoption, is rooted in a desire to rid the world of bad dating behavior. 

“We’ve got ghosting now, people don’t want to talk about what they really want, people don’t want to define relationships, people are playing the field and having three dates in one night,” Hosseini said. “If I could snap my fingers, I’d love something that cuts through all that.”

His wish sounds a lot like what Kang envisions. In 10 years, she wants this era of dating disappointment to be a distant memory.

“I hope we are thinking, ‘Oh wow, I can’t believe people used to swipe other people like that and just ghost people,” she said.

Us, too. 

More from Love App-tually

Uploads%252fvideo uploaders%252fdistribution thumb%252fimage%252f90359%252f0c85c165 4fb6 4b9f 930b 53621526912a.jpg%252foriginal.jpg?signature=1w0fshm1h2ypwjviam4kc1imwzw=&source=https%3a%2f%2fblueprint api production.s3.amazonaws

 

Continue Reading
Advertisement Find your dream job

Trending