https://www.wired.com/story/i-uncovered-an-army-of-fake-men-on-hinge/?utm_campaign=mb&utm_medium=newsletter&utm_source=morning_brew

In the land of love, there are fakes, and there are fakes. There’s the realization that the flesh-and-blood person you’ve spent time with is inauthentic in some way, the old-fashioned bluffing of the Homo sapiens mating game. And then there are the unnaturally smooth selfies and stilted messages that suggest an AI-generated facsimile of a person. On dating app Hinge, which claims to serve those seeking life-long connections, there appear to be a lot of these.

Hinge surely is not the only dating app riddled with digital fakes. The potential for romance makes people more vulnerable than in other digital contexts. In the early days of Tinder, people complained about chatbots that would encourage them to click on suspect game links. More recently, as usage of dating apps has soared during the pandemic, these services have been targeted by sophisticated social engineering operations known as pig-butchering scams.

I’m one of many millions who have bumbled their way through Bumble or taken a swing on Hinge over the past couple years. I never got into Tinder, because I just can’t, and that phrase alone should be evidence enough of my elder-millennialism, which explains why I never got into Tinder. I was invited to try Raya, and I’m sorry to report that I have absolutely zero sexy celebrity DMs to share with you. Hinge seemed to be the most straightforward of the bunch, until it wasn’t.

Over the past few months an increasing number of uncanny profiles have filled my Hinge feed. As I write this, I’ve just cycled through 15 profiles in the app—at least four have signs of being fake. Their photos are too polished, their profile descriptions totally nonsensical. For a while I ignored these oddities, partly because I wasn’t super invested in the app and partly because my brain has been coded to run scripts like:

If (abdomen.equals("Photoshopped")){ //DELETE MATCH ASAP };

But then I figured I’d “like” these bot profiles back, establishing a match, to see what I could dig up. Hinge itself soon confirmed what I suspected, booting some of these supposed people off the app for potentially fraudulent behavior. I received automated emails on August 6, 8, 9, 14, and again on September 18 and 22, letting me know a match was a fake.

Some tell-tale signs of a fake Hinge profile, if like me you are a cisgender woman seeking a man: These men have no blemishes, wrinkles, or quirky birthmarks. No amount of Retin-A and Instagram filtering would give me their complexions. Many profiles include a shirtless photo, and even the 40-somethings appear to have forsaken carbs, now and for the remainder of their days. Ai, whose name was a little on the nose, even chooses to go backpacking shirtless, nevermind the chafing.

They love to show off their dogs or any other cute animal that might draw your attention. Alex, 38, holds a baby lamb, while Pacheco dares to pet a couple of lion cubs and Matthew hangs out with a camel. Also like Matthew, they are often at the gym, or golfing, like Smith, or gardening, like Victor. It’s as if someone typed “Chris Evans playing with puppies” into some kind of Himbo AI generator and a million Hinge profiles were spit out.

These profiles frequently have a purple “Just joined” badge, indicating they’re new to the app. But it’s the text within the profiles that’s most awkward. Aaron writes “I bet you can’t spicy food.” Emi said he is looking for “A sincere, kind and caring Days and Nights in Wuhan,” a reference to a Chinese Communist Party documentary. Leon’s love language is to “use correct words and physical contact reactions.” Liwei is convinced that “Christianity.” That’s it, that’s the sentence: “I’m convinced that Christianity.”

Last month I contacted Match Group, the love leviathan that owns not only Hinge but also Tinder, OKCupid, Plenty of Fish, the video chat app Azar, the “high-standards” dating app The League, Match itself, and others. A company spokesperson, Justine Sacco, initially said she was surprised by my query, which surprised me, since the bot problem was so obvious. Later, another spokesperson, Kayla Whaling emailed a statement that said Match Group uses AI and machine learning to proactively ban bad accounts, and invests in “innovative technology and moderation tools to help prevent and disrupt potential online harms.” Somehow none of it had touched the swarm of bots that had contacted me.

Whaling also said that many of the company’s apps ask users to take profile photos within the app itself, so that automated tools can compare the images with the person’s already-uploaded photos. In theory, this provides evidence that a person is who they say they are. But this Photo Verification feature isn’t yet available on Hinge.

I’m not here often. I’m sorry. There’s no beep.

Match Group’s communications staff being little help, I decided to try conversing with the bots instead, hoping to understand how they work and what they’re supposed to accomplish.

A friend who works in machine learning suggested I lob random but highly specific questions at them, something like “What’s your favorite dinosaur?”, to try to trip up the chatbots. The first “man” I tried it on unmatched me soon after. Clearly I had caught a bot. Or maybe when you’re a grown woman you’re not supposed to ask potential dates “What’s your favorite dinosaur?”