Can an Algorithm Be Trusted With a Soul Mate Search?
WHY YOU SHOULD CARE
Artificial intelligence still has to be designed by people.
By Carly Stern
On one end of the line, Kevin Teman told her he was going to end it. But the woman on the other end wasn’t his girlfriend. She was his matchmaker.
Tens of thousands of dollars and one year spent telling another human what he wanted in a partner had led to 10 introductions. Teman felt listened to and cared for throughout the process — which was the complete opposite of his experience using dating apps. But none of the introductions resulted in sparks.
And so the software developer’s disappointing love life led him to a counterintuitive conclusion: To get the more human dating experience he craved, Teman would build a machine to help him.
This is what I came to discuss with Teman in his apartment with a view of the Denver skyline on a September afternoon. We’re chatting at his kitchen table, next to countertops surprisingly cluttered for a 36-year-old engineer who prides himself on efficiency. Flyers with phrases like “Make dating great again” and “I am always turned on” rest on the table.
The flyers are advertising materials for AIMM — Teman’s acronym for the world’s first “artificially intelligent matchmaker.” Launched in 2017, the app is an intriguing addition to the almost comically oversaturated market of dating apps. AIMM is leading us, ready or not, into the era when robots set us up with our future romantic partners. And that pitch is working, to a degree: AIMM has 1,300 users in Denver, Los Angeles and New York, and is soon expanding to other cities.
So how, exactly, does an algorithm find your soul mate? The platform, which Teman says currently is free, begins by asking users (in a British-accented female voice) 10 minutes of questions each day for a week. Unlike Apple’s Siri or Amazon’s Alexa, Teman points out, AIMM has voice recognition, is fully conversational and even cracks jokes.
After crunching the data, AIMM matches one would-be partner with another, sharing photos and audio clips before facilitating a phone call. Partners don’t exchange messages before the first date and the platform follows up postdate to debrief. Didn’t feel any chemistry? No need to ghost and feel guilty about it; AIMM will politely cut ties for you — “like if you had a recruiter for a job,” Teman says.
That may sound harsh. But heading into this interview, I could understand aspects of the app’s appeal: AIMM could help busy (or anxious) people for whom the predate process can be a paralyzing burden. Perhaps automating some processes could leave space to personalize others. And while the philosophy of most modern dating apps is “honey to bees” — attract a high number of prospects and eventually you’ll find a good fit, as Bryan Harris, a video producer who has worked for AIMM, puts it — this is a targeted, methodical approach to finding the one person you’re supposed to be with. Because the algorithm does legwork before the date, you can focus on getting off the service quicker and living in real life.
For all the modern idealism when it comes to love, Teman himself has some troublingly antiquated views. “One person is basically the pursuer, and one person is the pursuee,” he says, noting how in AIMM’s world, only men are pursuers — though he intends to work in a feature so users can pick which role they prefer. Men are asked certain financial questions that women aren’t. “In a traditional relationship, a lot of times women want a man who earns a good living,” Teman says. That said, roughly half of AIMM’s employees are women, and the company has worked with female testers and matchmakers. Five percent of the app’s users are LGBTQ.
Men are given two to four choices to pick from, while women are given one at a time. What’s more, women have been encouraged to stick with a match they have hesitations about. When I ask about challenges people have reported with the app, Teman says some women “wanted an out” so they could “reject the guy earlier” when they didn’t like AIMM’s pick. “AIMM was saying, like, ‘Hold on a second. Wait for a while so I can show you a few more things.’” He adds, “People need to have a little more patience with the process,” although he eventually altered the app to accommodate those women’s concerns.
Teman is gangly, polished and slightly awkward at times. He wears a business-casual outfit: blue button-down, linen pants and black dress socks. The son of Southern California engineers who split up when he was young, Teman was raised by his mother with “logic at the expense of emotion,” he says. He has a business degree — he also holds a second-degree black belt in karate — and has been a developer for roughly 10 years.
When a three-year relationship ended in 2012, Teman began again. He describes himself as a lone wolf and says he tried to meet women through “all the natural things” but struggled. “I tried meeting people at grocery stores,” he says. (I, for one, would find it invasive to be hit on while buying eggs.)
He had no luck on Match.com, OkCupid or Bumble, even though he tried to do “all the things you’re supposed to do to meet someone.” That included hiring a photographer to take professional headshots and getting help from others to write his profile.
For Teman, apps provide and commodify the hope of meeting someone but do little to facilitate it, instead forcing users to be their own press agents. Given the way social media has hijacked how people view themselves — with upticks in depression and anxiety — having unbiased representation could be welcome. Of course, that’s assuming people are ready for an algorithm to hold the mirror rather than curating their own images.
Females always have an easier time on most dating apps.
AIMM promises to smooth the transition from stranger to partner; in effect, to offer both the training wheels and the guided path. Anything that gets people off apps and into real life sooner marks an improvement, says Helen Fisher, chief scientific adviser for Match.com. “The only real algorithm is your own brain,” she says.
The platform also coaches users before first dates and offers confidence boosts. Isn’t that what friends are for, though? Teman points out that friends, while well-intentioned, can often be less plugged in. For users, simple directions can be more comforting. “If I want to go on a diet or something, I need to be told what to eat. I don’t need all the choices,” says Beth Sauer, a 45-year-old who briefly used AIMM after a long relationship ended. The questions went deeper than those on other apps, she says, although she wasn’t interested in AIMM’s match.
It’s easy to do the wrong thing, says Teman. He’s made those missteps. “Something I had done wrong many times before in my dating life is coming on too strong too fast because I’m an intense person,” he says.
Why not let a machine make the calls rather than put yourself out there when intuition could lead you astray?
While Teman comes off like he genuinely wants to help people, he also shows moments of entitlement. He says he felt “kind of ousted that I didn’t find my person” through traditional matchmaking given the service’s high success rate — although his matchmaker worked hard to help him and was “apologetic,” he says. “She was really, truly sad that she couldn’t find the one for me.”
“Females always have an easier time on most dating apps,” he argues. “Guys don’t have time to send the messages required to get responses.” He’s crystal clear about what he needs: “Somebody who brings as much to the table in my relationship as I bring to the table.”
All of which makes me wonder if the man behind the algorithm pairing women with their supposed perfect match understands women at all. Teman’s description of AIMM seems less about convenience and more about eliminating the emotional work required to understand oneself and others. Attraction can’t be debugged like code. Everyone hopes investing energy leads to outcomes. But paying a fee for a service, like Teman did, doesn’t mean customers are owed the relationships they seek.
AIMM matters more than one person’s love life. Already, artificial intelligence expected to eliminate bias — in everything from hiring practices to policing — has proved to perpetuate it. As algorithms increasingly dictate what we see, do and think, the people shaping those equations have tremendous power to shape our lives as well. Even when it comes to love.
To this process, Teman brings optimism. “I believe I will end up with somebody that really makes me fly,” he says. “And I know it’s there.” But he also brings assumptions, the same ones that will drive the experience of every user on his platform.
In the world of AIMM that Teman has created, he’s accustomed to setting the rules. “I’m going to interview you after this,” he jokes. I laugh but don’t oblige.