Case in point: , a new app which uses neural network-based facial recognition AI to scour dating platforms for profiles that match a user’s specified ‘type’–or even belong to a specific person. As with many of our great cyber ‘success stories,’ the concept is a simple one–using extant technology to find the faces we’d most like to caress (or whatever)–but fails to account for a wide range of complicated, very human concerns, touching on everything from online privacy to real-life values.
In the app’s free version, users can choose from images of around 200 regular and online celebrities and perform an unofficial crawl of platforms like Tinder, OkCupid, and Plenty of Fish to find doppelgangers located anywhere. In its premium version (available as a one-week trial or for $7.99/month), users can choose to Dating eines alleinerziehenden Vaters tailor their search for the perfect face by gender, age, zip code, and by uploading as many pictures as they like, acquired however they like, of whomever they’d like–say, an ideal-seeming actress, founder Heath Ahrens suggested, or even an ex-girlfriend.
For generations, technology has been driving the way we date; with both innovation and relationships, however, it is vitally important to realize when it’s time to hit the brakes
In a phone interview last week, Ahrens explained that the app is just as useful for finding lookalikes to past lovers as it is for finding faux celebrity crushes. This innovative method of locating partners, which officially displays only «user-generated content» from its robo-crawls through big dating sites (shifting liability to its users), has been a long time coming, and is something the online dating scene needs now, according to Ahrens.
«There just hasn’t been enough advancement in dating apps–Tinder was a big advancement, years ago, but there’s been nothing really since,» he said. «There’s been a need for something that can swipe for you, put you right in front of the people you’re interested in. Say my relationship didn’t work out, but I was really attracted to that person; so, let’s upload a picture of my ex and find someone who looks just like them but isn’t as mean and nasty, who will listen to me, be kind, and so on.»
«Users are finding people who look just like their exes, and like celebrities; so far, it’s working really well,» Ahrens said.
After segueing from speech recognition technology to facial recognition a few years ago, Ahrens decided to put his new company to work toward getting a technology that’s already deployed by Facebook and Google into the hands of individual users.
«W e had a few single guys in the office who came up with this idea, but at the same time, we’re trying to showcase technology that could be used for all different kinds of things that we haven’t envisioned yet, and lots of smart developers will come up with other uses,» he said. According to Ahrens, users have already reported uploading their own photos to see whether they’ve been ‘catfished’ (meaning someone’s ripped off and reused their photos) and to determine what they’d look like as another gender, based on the AI’s unseen but growing idea of who looks like what, and how chromosomes might tie in.
Ahrens noted that helping users finding visually curated sexy-friends wasn’t his only goal with the app
As Ahrens explained, the app leverages a developer API called Haystack, which allows developers to use the technology for different applications. The «AI» in arises from a specific neural network, or self-teaching type (rather than one with developer-framed algorithms), which does a lot more-data crunching in its image quest than occurs, for example, in a Google text search. According to Ahrens, a search engine for faces rather than for terms like «surfer girl» is really all that his app is, albeit somewhat slower.