An Incredibly Accurate Facial Recognition App Is Coming — Here's What It Means for Privacy


Privacy is dead — or at least, it will be soon. That's the conviction held by Russian entrepreneurs Artem Kukharenko and Alexander Kabakov, whose startup, NTechLab, recently launched a facial recognition app that nearly obliterates the concept of anonymity. Called FindFace, the app has remained exclusive to Russia since going live earlier this year. Soon, though, Kuhkarenko and Kabakov are introducing a cloud-based platform that makes their frighteningly accurate algorithm available to everyone, the pair said in a Skype interview in May. 

FindFace indexes photos from VK, Russia's Facebook equivalent, matching them to uploaded photos with 70% accuracy, according to Kukharenko and Kabakov. In a practical sense, what this means is that none of us is safe from an always-probing public eye.

"In 10 or 20 years, there won't be a place on the earth [where] ... nobody [can] see you," Kabakov said over Skype. This is because cameras come attached to pretty much every gadget available — phones, laptops, televisions, refrigerators — and because our digital fingerprints are all over each of our social media interactions. 

As Russian artist Yegor Tsvetkov illustrated all too vividly with his FindFace-based project, our faces have become big data. Using his phone to covertly photograph some 100 strangers on Moscow trains, Tsvetkov then successfully located about 70 of his subjects through FindFace. 

"My project is a clear illustration of the future that awaits us if we continue to disclose as much about ourselves on the internet as we do now," he said of the project. In other words, a future without privacy.

How FindFace works

It seems people are all too eager to test-drive the technology themselves. In an emailed statement from April, NTechLab claimed its software has been downloaded over 250,000 times, but the Guardian more recently reported that the number had surged to 500,000. Something like 46 million people use VK every day, while 100 million active profiles exist on the site. That's compared to Facebook's 1.09 billion active users in March 2016 alone, and 989 million daily users. According to the Guardian, FindFace can't index Facebook photos because they're stored more securely than are VK photos. But theoretically, any public images are free search game. 

In 2015, NTechLab won the MegaFace challenge at the University of Washington, its facial recognition algorithm besting Google in terms of accuracy. NTech then launched that algorithm as FindFace, pitched as a dating app that allows users to track down desirable strangers by uploading a photo of that stranger's face. 

According to Kukharenko and Kabakov, that photo would be compared to a pool of 300 million pictures within VK's database, which the algorithm indexes almost immediately on just five servers. It's basically a search engine for faces. 

So if, for example, you spied an alluring beauty lurking by the train doors during your morning commute, you could sneakily snap her photo on your phone and feed it to FindFace. The app would furnish about 10 VK profiles of individuals who look like the stranger, which means you'd have access to these individuals' names and contact information. If the stranger you're seeking isn't on VK, FindFace will still show those 10 similar-looking individuals in what is essentially the human version of Netflix's "recommended for you" section. 

What FindFace means

Hear that? It's the sound of millions of women groaning at the prospect of yet more channels for anonymous individuals to make unwanted advances and/or harass them. Many of the possibilities this hyper-accurate facial recognition software introduces are undeniably creepy, especially in a romantic context.

But the inventors claim that FindFace isn't a dating app, per se. Rather, they said it serves to display "the power of our technology," which they see as being more useful for public safety. Some 10 police departments, they said, have written in to say FindFace had allowed them to obtain long-sought-after perps and solve cases. It's also apparently helped parents to locate runaway teens and estranged classmates to reunite, according to Kukharenko and Kabakov, who believe FindFace holds promise for airports, public transportation, border control and other broad-scale security initiatives. 

The pair said they have been approached by international government agencies and are currently in talks with Moscow authorities who are seeking to employ their technology in 150,000 surveillance cameras around the city.

Kukharenko and Kabakov also spoke at length about FindFace's potential to revolutionize the retail experience both in-person and online. A customer enters a store, the security cameras capture their face and, Kabakov explains, "you can see every, any statistic about your customers in real time." Immediate access to all conceivable consumer data would help companies to devise better loyalty programs and connect with customers in unprecedented ways.

In theory, all the above sounds anywhere between fine and great — but do more aggressive rewards programs or more effective police surveillance systems justify the end of privacy? 

Can't fight progress

The app's creators maintain that no one can — or should — stop the forward march of technology. Facial recognition software isn't confined to Russia: Facebook has a program called DeepFace, which the company hopes will be able to recognize a face with "human-level" accuracy, and North American retailers are already experimenting with precisely the kind of customer tracking Kukharenko and Kabakov describe. 

And then, the pair have plans to launch a cloud-based FindFace platform this summer, which they said could be used by "any company, all companies" for their facial recognition purposes. 

The FindFace creators suggest that users who don't want their faces found can "set their permissions to private," but as Jonathan Frankle, staff technologist at Georgetown Law's Center on Privacy & Technology, pointed out in a recent Atlantic op-ed, the only way a VK user can make their profile picture private is to block a specific user. Making one's face unsearchable, then, means just not having a VK profile. As long as the technology remains within the realm of VK, that is. 

But as cameras become ubiquitous, Kabakov said, transparency is what's important: that people know their lives are public, and that they think actively about what they post, for whom they post and which words they use. Technology like FindFace, Kabakov said, might make for a world in which people are more conscientious because they know they're being watched. 

It's easy to see where it could just as easily make for a world in which harassment is rampant because anonymity is obsolete, no longer a right. We'll find out either way, because as the pair stressed, this future is inevitable.

"Prepare for the things you do to be public," Kabakov said. In other words, prepare to live in a world without privacy.