Commentary: AI dating might actually not be as bad as we imagine

NEW YORK: Most people seem to instinctively reject any suggestion that artificial intelligence might be able to help people find love. When Whitney Wolfe Herd, the founder of dating app Bumble, said earlier this year that AI concierges might one day “date” for us, the idea was called a “dystopian hell” and “the end of the road for humans.”

I was in the audience for Herd’s comments, and I’ll admit I winced along with everyone else. But now that I’ve had a little more time to think about it, I think we need to give the prospect of AI-enhanced dating some additional consideration. 

Let’s start with the problem. When people talk about “dating app fatigue” – as they do a lot these days – they’re often referring to the limitless stack of profiles, arranged by a mostly dumb algorithm, awaiting an instant and superficial “yes” or “no.”

An empty message box then calls for something more compelling than “how was your Monday?” or “so how long have you lived in New York?” (The only thing more tedious than sending one of these is having to reply to one.)

People have also grown weary from heading out on first dates, optimistic from some good early text exchanges, only to find that the person isn’t what they were hoping for. The connection, for whatever reason, just isn’t there.

But this too is partly the fault of the app, a technology fundamentally limited by the fact software code can only measure matters of compatibility – age, location, job – rather than understand the more subtle signs of potential chemistry. Centuries of literature and love songs have struggled with this, too.

FINDING THE PERFECT MATCH ON DATING APPS

Maybe AI can help do some demystifying. By analysing more data, apps could understand our preferences better than we might understand them ourselves, helping us hone in on the perfect match.

It could be an innovation as disruptive to dating as the introduction of Tinder’s swipe in 2012, suggests Justin McLeod, chief executive officer of Hinge, an app owned by the Match Group.

“It requires getting a level or two deeper than we currently do,” he said. “There is a lot of latent signal in photos,” for example. “It’s not just looks. There’s all kind of information that gets portrayed – how someone carries themselves, what kind of photos they choose.”

Tinder’s appeal when it arrived 15 years ago was the introduction of swiping left or right on potential suitors, making online dating easier and drastically increasing the dating pool. Competitors offered similar ease with their own spin: Bumble made it so only women could send the first message, while Hinge required users to interact with a specific part of a person’s profile rather than making a split-second decision. 

Yet, over time, what seemed liberating also brought on a feeling of hollowness. With so many options, people began to feel expendable. Words like “ghosting” entered the lexicon. The successful dating apps of the future will take the same broad field but narrow an individual’s choices back down again, giving more meaning and consideration to each.

“If we don’t figure it out, someone will,” McLeod said. The goal is “to move from a world where you’re voting up or down on dozens or hundreds of profiles to a world where you’re providing really thoughtful feedback on just a few. I think we’re underestimating how different that’s going to feel for people”.

CHANGING THE WAY PEOPLE MEET ONLINE

All this might be wishful thinking when the dating industry is looking eagerly for new ways to grow.

There are concerns Gen Z isn’t as keen to date online as the millennials were. McLeod said he doesn’t think those dynamics apply to Hinge, whose year-on-year revenue growth of 48 per cent in the last quarter far outpaces that of Tinder (1.1 per cent) and Bumble (4.8 per cent).

Gen Z, McLeod said, is Hinge’s fastest-growing cohort. Still, as part of the Match Group, the app is under the glare of no fewer than three activist investors that took stakes in the parent company over the past year and are now pushing for cost efficiencies, smarter monetisation and user growth. (McLeod said he hasn’t talked to any of the activists yet.)

Hype or not, it’s clear that over the next few months daters will be subjected to an overhaul in how they meet people online thanks to AI. The plan is to sell it to users as a way to not just find better matches but to become more appealing to potential suitors as well.

This doesn’t mean fakery – the company has protections in place to prevent AI-generated or enhanced photos – but highly personalised advice and guidance.

LESS DYSTOPIAN THAN AI DOING OUR DATING FOR US

McLeod said he has been speaking with old-school matchmakers to learn more about that craft. "Their primary job is actually the coaching and helping their clients," he said, adding that he thinks Hinge "will start to feel closer to the experience of having an expensive, personal matchmaker focused on you, helping you really zero in on the right person".

The company is already testing a system that will tell users if their prompt responses – the little bits of text on a profile designed to entice matches – could be improved. One trap users fall into, McLeod said, is offering generic details (“I love brunch!”) that don’t give much opportunity for an interesting (or flirty) response.

“We’ve trained models on responses that actually yield conversations and dates, and those that don’t,” he says. “We can look at that and say: ‘Hey, can you tell us a little bit more about brunch? Where, with whom, and what do you talk about? And why is that important to you?’ We found it’s actually really effective. It’s not that people don’t want to put in the effort. They’re just a little scared. They don’t know what to say.”

A helping hand to sound more interesting when talking about brunch feels far less dystopian than Herd’s vision of an AI concierge doing our dating for us.

But it’s ultimately about the same aim: Using AI to present the best version of yourself to people you’re more likely to actually be interested in. Rather than lumping you in with the tastes of everyone else, AI can do a better job of understanding your personal preferences.

After all, you’re not looking for the person the average Hinge or Bumble user finds hot, you’re looking for the person you find hot, for whatever reason that hotness floats your charmingly unique kind of boat. That certainly sounds like a healthier, more productive way to date.