advertisement
advertisement
  • 08.07.15

Has This Man Unlocked The Secret To Internet Anonymity?

Jonathan Hirshon kept his photo off the Internet for 20 years. Now he’s rallying the troops to hack Facebook and Google facial recognition.

I’ve met Jonathan Hirshon. I know what Jonathan Hirshon looks like. You probably don’t, and neither does Google or Facebook. That’s just how the PR professional wants it.

advertisement

Hirshon is shorter than most, and balding. He wears glasses. He’s friendly, and well liked in the tech industry. Back in the late ’90s, he did in-house PR for Apple.

At 48, Hirshon was an adult well before the Internet was a mainstream tool. Though he’s a pro when it comes to getting his clients publicity, he’s dead-set against his own image being online.

It must have worked, since a Google image search for his name doesn’t return a single picture that’s actually him.

“When the [web] started, I decided to play a game with myself: How long could I keep my picture off the Internet,” Hirshon told me. “It’s turned into a 20-year trek….There’s only been two instances where pictures were taken of me [and posted] without my permission, and they both took them down.”

This might all strike some as odd, particularly because, as he says, “I live on Facebook.”

Recently, Hirshon admitted to himself that his assiduous control over his image could soon slip from his grasp. Cameras are everywhere, after all, and you never know who’s Instagramming.

advertisement

It was time for some culture jamming.

Throwing Down The Gauntlet

On August 1, Hirshon tossed a wrench into the gears of Facebook and Google facial-recognition algorithms built to threaten his 20 years of “cherished anonymity.” In a post on Facebook—visible only to his friends, or course—he threw down the gauntlet.

“If you’re so inclined,” he wrote, “take a moment and tag me in some random picture or image. A leaf on the wind, a howler monkey, geometry equations, George Clooney, a large steaming pile of excrement—select an image that you think best suits me or [is] based solely on your whim.”

The purpose was clear. If enough people did his bidding, the collection of false positives associated with the tag “Jonathan Hirshon” could “both confuse Facebook and Google’s algorithms and help to bury any real pics of me that eventually turn up in a sea of anarchic image randomness.”

By phone, Hirshon called his challenge “kind of a combination of psychological and sociological hacking, all mixed together. Or algorithmic hacking.”

Should you wonder how committed Hirshon is to remaining visually anonymous online, I offer this evidence: In May of 2013, after a meeting with tech superstar Robert Scoble, who at the time was wearing Google Glass 24 hours a day, Scoble posted a photo of Hirshon, with the following caption: “Hah, PR dude Jonathan Hirshon doesn’t want his photo on the Internet so he showed up today with a device to foil my Google Glass.”

advertisement

That decidedly low-tech device was a balaclava. And I don’t mean the Greek dessert.

Or take the 2014 video posted by The Unofficial Apple Weblog (TUAW) about Hirshon’s years working at Apple. Throughout the 11-plus minute interview, he appears only in profile, in the dark, almost like he’s in witness protection.

You get the picture.

Witness protection?

Why does he care? “If I told you that, I’d have to kill you.”

The real answer, though, isn’t much more revealing.

“You’ve probably noticed that I have a fairly decent grasp of facial recognition technology,” he told me, and “I have worked for a number of companies in the (information and security) space. You can leave it at that, and let people draw their own conclusions.”

advertisement

No Canonical Image

Hirshon’s new experiment has energized many of his Facebook friends. More than 20 people contributed dozens of photos, all tagged with his name. They ranged from a pig pushing a shopping cart to a Buddha chill-out poster to Bill Nye the Science Guy to Don Draper to a Masonic logo, Brad Pitt, a cat, Abraham Lincoln, a seagull, and so on.

My personal favorite: Dos Equis’ Most Interesting Man in the World with the caption: “I don’t always go by the name Jonathan Hirshon.”

“When he mentioned this experiment, I thought it was brilliant,” said Elizabeth Roman, a Hirshon friend who contributed more than 10 different photos. “I can see how having a ton of images come up for any particular name can make it impossible to know for sure which is accurate. I chose images that represent his personality, hobbies, and interests plus a couple funny ones, like Nosferatu (that one also looks the most like him).”

Going through the full stream of pictures, it seemed like a lot, but it wasn’t really—a few dozen at best. And for most people, who have dozens or hundreds of photos of themselves online, as I do, this would hardly be enough to monkey with algorithms designed to identify specific signals amidst floods of interference.

“Computers are dumb,” said Declan McCullagh, the cofounder and CEO of Recent Media, an AI-powered news app startup, “but they’re very good at filtering out noise.”

Hirshon believes he’s one of the few who can disappear into the noise and that he’s the perfect test case for a hack like this because there’s no baseline for Google’s or Facebook’s facial-recognition algorithms to work with.

advertisement

What he’s really saying is that neither Google nor Facebook have a “canonical image” of him, one that establishes the shape of his nose or lips, the distance between his eyes, or other facial features their algorithms can use to hunt for matches.

Even better, Hirshon insists, the new experiment isn’t just about him.

By tagging pictures of themselves with Hirshon’s name, he argued, other people can gum the works and begin to reclaim some small bit of their own privacy. It’s “like crowdsourcing anonymity. If enough people do this, it will get to the point where facial recognition for the average person will become not impossible, but a lot more difficult.”

True? Perhaps, but more likely wishful thinking.

“It’s a clever idea, and it may work,” said McCullagh, who wrote frequently about security and privacy as a technology journalist. “On the other hand, algorithms are getting smarter, so I’d be surprised if it’s a long-term solution. If there’s a cluster of photos that’s persistently one person over months or years, that may be enough to weed out the randomness. And smarter algorithms will [give more weight to] photographs from authoritative sources, like news articles or corporate bios. The difficulty of staying photo-anonymous online is it just takes one slipup, and you’re up against increasingly intelligent algorithms that will pounce on it.”

Neither Google nor Facebook commented for this story.

advertisement

Liking Everything

This is, of course, hardly the first time someone has used a giant tech company’s own tools to try to confuse it.

Last year, then-Wired reporter Mat Honan spent 48 hours liking every single thing he encountered on Facebook in order to circumvent the algorithms the social network uses to determine what appears in our news feeds.

“My News Feed took on an entirely new character in a surprisingly short amount of time,” Honan wrote in an article that went viral. “After checking in and liking a bunch of stuff over the course of an hour, there were no human beings in my feed anymore. It became about brands and messaging, rather than humans with messages.”

More recently, artist Kirk Kaiser decided to run numerous photos of himself through Google’s artificial neural network system, known as DeepDream. The result was deeply modified images he then posted to Facebook.

“Every time somebody takes a photo of you or you upload a photo of yourself, it’s added to the training data set that exists in the ether of who and what you are,” Kaiser told Popular Science. “The general idea is to corrupt the data set that exists on us and get back a little bit of control.”

That’s very much what Hirshon has in mind, even if his methods are quite different.

advertisement

Hirshon has a sense of humor, and he’s calling his experiment the “Spartacus hack.” That’s based on the 1960 Kirk Douglas film in which hundreds of slaves are offered mercy if they will only tell the Romans who among them is Spartacus. As Douglas stands up and says “I’m Spartacus,” dozens, and then most of the hundreds of other slaves quickly rise to their own feet, shouting, “I’m Spartacus.” The Romans were deeply confused.

Hirshon calls his experiment “the Spartacus hack.”

Perhaps Facebook and Google are smarter than the Romans. It’s hard to say right now, given all the false positives in an image search: former NBC News anchor Brian Williams, a dogsled team, Redef.com CEO Jason Hirschhorn (bizarrely referred to as Jonathan Hirshon in a PandoDaily video), and more.

For now, then, it’s working. Hirshon is defiant.

“It’s basically a way to put your fist in the air, and say, ‘I want to make a difference,'” he says. “‘I want to reclaim a small measure of my digital identity.’”

I count myself as privileged for having met him. I can pick him out of a crowd. But even if I have the opportunity, I probably won’t be the one to ruin his multiple-decade-long project to keep his face off the Internet.

“That’s the one thing I still hold very close, and it’s amazing the reaction people get when they finally meet me,” he told me. “It becomes like an event. Like, ‘I’ve seen Jonathan Hirshon, wow!’ People collect my cards. It’s like seeing Bigfoot.”

About the author

Daniel Terdiman is a San Francisco-based technology journalist with nearly 20 years of experience. A veteran of CNET and VentureBeat, Daniel has also written for Wired, The New York Times, Time, and many other publications.

More

Video