What emerged from Mark Zuckerberg’s rather crude ramblings that night was:
a privacy reprimand by Harvard officials
a website called Facemash that became Facebook
a company with a book value over fifteen billion dollars
a communications system that has 1.3 billion active users just about everywhere on the planet, some of whom have to defy their repressive national governments to sneak in a posting or two
I can assure you that Mark Zuckerberg did not create the fundamental concept behind a face book. I know this because I am holding a dusty “1970 Freshman Directory” from Columbia University. Long before Mr. Zuckerberg was even imagined, college students were already judging each other and ridiculing awkward high school grad photos, in dorms and dining halls across the country.
One of the Columbia College men would trade a copy of our all-male book with a student from the all-female Barnard College just across Broadway, and we would engage in the same late-night “hot-or-not” discussions that Zuckerberg automated. However, things posted on Facebook now travel instantaneously around the world, which vastly increases the potential impact on our lives.
In 2010, Facebook introduced a new feature to automatically tag people in photos through facial recognition. This “tag suggestion” feature was turned on by default, a situation that did not sit well with data protection authorities, especially in Europe. The people there seem to be a lot keener on privacy protection than many other nationalities: images from the 1940s of punch cards with meticulously-typed Jewish names are etched into the public’s consciousness. Remember that the vast majority of Facebook users provide their real names and photos, bowing to the company’s Terms of Service.
When data protection commissioners in both Ireland and Hamburg objected to automated facial recognition, Facebook removed tag suggestions from customers in those countries.92 In fact, they even removed the feature for users in the U.S. for a while, though it has been brought back in substantially the same form.93 The company is coy about the exact number of photos that are in its database, but did say in an SEC filing that “on average, more than 250 million photos per day were uploaded to Facebook in the three months ended December 31, 2011.”94 So, Facebook gets to build the world’s largest, self-validated photo database on the planet, a project which has mind-boggling value for everyone from marketers to dictators to law enforcement agencies.
Not content to rely on careless, lazy humans to properly tag photos with names, Facebook’s Artificial Intelligence Group in Menlo Park, CA has been hard at work on “DeepFace: Closing the Gap to Human-Level Performance in Face Verification.” In a 2014 academic paper they reported that their method “reaches an accuracy of 97.25% on the Labeled Faces in the Wild (LFW) dataset.”95 LFW, maintained at the University of Massachusetts, is a popular collection of more than 13,000 faces with names attached that is often used for testing facial recognition technology.96 This amazing performance rivals that of humans, who, we are told, are only about one quarter of a percent better (97.53%) than the algorithm. The scientists accomplished this by training a neural network “on an identity labeled database of four million facial images and by applying 3D rotations to align images.”
The announcement of DeepFace was greeted with headlines such as “Just as Creepy as It Sounds”97 and “Facebook’s Freaky DeepFace Program Knows Your Friends Better Than You Do.”98 The technology is still on the drawing board but it is hard to imagine it will not move into mainstream use quickly, perhaps even built in to your next smartphone.
The Onion News Network has a tongue-in-cheek video report claiming that Facebook is actually a “massive online surveillance program run by the CIA.” It goes on to say that “Facebook has replaced almost every other CIA information gathering program since it was launched in 2004.” The report praises “CIA Agent Mark Zuckerberg, who runs the day to day Facebook operation for the agency.” It jokes that Facebook’s Calendar feature even shows where you will be in advance so “now if they want to pick you up for questioning, all they have to do is see which events you’ve RSVPd ‘Yes’ to.”99
In the same vein, a conspiracy theory video on YouTube called “Does what happens in the Facebook stay in the Facebook?” tracks some of the early backers and funders of Facebook, highlighting their defense and intelligence community connections.100
I once put the question “Did the CIA create the Facebook” to someone senior enough in that agency to have a well-informed opinion. “No, we did not,” he said, but then he added that they use it every day as an excellent source of intelligence, and if it had not been launched by Zuckerberg, the CIA might well have created something like it.
MIT professor and author Sherry Turkle explains why we have such a burning desire to share our lives with the online world, including total strangers, in her book Alone Together. Turkle tells us that interacting with machines “may offer the illusion of companionship without the demands of friendship.”101 Although she is writing about robots, the point applies to Facebook as well. Although your ultimate goal may be to communicate with other human beings, when you add something to your Facebook you are in fact dealing with a technological entity.
Blogger Maggie Clayman expands on that idea, suggesting that “perhaps people share pictures of them[selves] with their children because they feel a need to prove that they spend enough time with their children. Perhaps people post pictures of new purchases or great meals because they want to prove that their lives are really good.”102
Photos of our favorite meals and recent purchases have marketers licking their chops. Analysis of that data, combined with images in our postings, can easily reveal other things that we might want to buy. To illustrate how this might work, consider a photo of loved ones that I posted, and tagged, on Facebook.
In the very near future, some computer will probably analyze this snapshot and come back with: “Hmm, (tag: Keri) that poor dog’s (tag: Joey) leash is looking rather tattered—we just happen to have a sale on them, running through today only. And those boots you seem to like, we’ve got them too. By the way, that ski resort you seem to enjoy still has space available over the Christmas holidays—if you book now. Click here. Oh, and are you interested in the paranormal? (tag: UFO) We have books on that.”
There are some entertaining countermeasures that can be deployed to befuddle the bots trying to analyze your images. I often tag delicious items on my dinner plate with the names of people I know, and Facebook plays along, at least for now. However, soon it will probably ask “don’t you mean Market East Coast Oysters on the Half Shell?” and offer to have more sent to me via an online food delivery company.
Even though most computer users understand at least some of the implications of posting and tagging photos, they seem to have decided, either consciously or implicitly, to go along with the game. Privacy experts disparagingly called folks like this “sheeple.” They are also the ones who give up their email addresses for a free magazine subscription or answer a detailed online survey, hoping to win a $500 gift card. They make a Faustian bargain with online companies, allowing total access to their lives in exchange for services that appear to be completely free.
With improving technology, once even a single good photo of you is tagged with your real name on Facebook, your privacy is a goner. You will be identifiable and trackable—unless you are prepared for a face transplant, or at least to radically modify your appearance.
Realface Glamoflage shirts were designed to do exactly that. Artist Simone C. Niquille has created multi-face designs to distract and confuse facial recognition software. Her shirts have a number of faces, including that of Michael Jackson and Barack Obama, and, for now at least, seem to confuse the face bots.
An even more radical approach to dodging facial recognition cameras has been suggested by artist and researcher Zach Blas, with an idea he calls Facial Weaponization. He helps people create masks with weirdly-morphed versions of their actual face, hoping to bedevil the rec
ognition software.
Figure 6. Zach Blas: Facial Weaponization Suite: Fag Face Mask–October 20, 2012, Los Angeles, CA. Courtesy of Christopher O’Leary.
Wearing masks in public can get you in trouble with the law in some places, but surely they cannot ban outlandish hair styles and creative makeup. Artist Adam Harvey has been experimenting with eye-catching patterns that put the facial recognition programs off your digital scent. He got his inspiration from Dazzle, a camouflage paint used on battleships in World War I.103 In the same spirit, photographer Petr Prokop has created Face Dazzler, a smartphones app that distorts your face in photos. He claims they become immune to facial recognition programs, but are still recognizable to your friends and family.104
There are over 70 synonyms for “friend” in Roget’s Thesaurus—“acquaintance, neighbor, well-wisher, advocate”—but none of them is “a person or entity you don’t really know but who seems to share your prejudices so you clicked yes on their friend request.” All technologies, and especially social media ones, expropriate familiar words and create their own creepy vocabularies. What is a poke? A news feed? A timeline? A like? Surely not the same as in the real world.
In a fascinating experiment, Dean Terry and Bradley Griffith at the University of Texas, Dallas created the concept of a Facebook enemy. Using their EnemyGraph application, users can declare their undying hatred of a politician, musician, or habit like smoking and see how many others chose the same enemy. Justin Bieber makes the list twice because he has multiple incarnations on Facebook.
For a while, whimsical entries such as “truck balls” and “bunnies” made the “Top Enemies” list, but as the EnemyGraph app attracted more users, the true enemies of the people rose to the top. Facebook normally bans applications that could disrupt the monetization of their ever-growing network of connected human beings. So if anyone tries to launch an “unfriend everybody” functionality it is quickly shot down by the company.
Yet Facebook tolerated EnemyGraph, perhaps because someone there understood that it was unwrapping a whole new category of information. Users could now be grouped by their mutual dislike of Bieber or bunnies or Microsoft, so targeted ads for things like bunny eradication powder might be both possible and probable.
Terry noted on his website that “people are also connected and motivated by things they dislike. Alliances are created, conversations are generated, friendships are stressed, stretched, and/or enhanced.”105
While not ignoring the commercial possibilities of the data he is generating, he says it is really just a fun social experiment and critique of the philosophy of Facebook: “So, Facebook runs queries to find affinities. EnemyGraph runs what we call dissonance queries. So if you have said you like, say, Portlandia (a TV show) on your profile page, and in our app one of your friends has declared them an ‘enemy,’ we will post this ‘dissonance report’ in the app. In other words we point out a difference you have with a friend and offer it up for conversation, as opposed to a similarity. Relationships always include differences, and often these differences are a critical part of the fabric of a friendship.” He goes on to suggest that in the “country club atmosphere” of Facebook these differences are ignored because dissonance is not part of their “social philosophy.”
Some are even suggesting that the digital trail you leave simply by “Liking” things on Facebook can paint a fairly accurate profile of you. The website www.youarewhatyoulike.com tells you about yourself based on your Facebook likes. I dug into the logic of the program using a specially constructed new Facebook profile to see which likes it used to form its opinion. It reported two of the ten Likes on the profile, and they were both travel sites. It told me pleasant things like I am “warm and trusting” and “liberal and artistic.”
The algorithm behind this comes from a research project at the University of Cambridge. For marketers, this kind of application is real gold. You can be sure if this level of analytics is available to all of us, for free, what they have is a lot more powerful.
While this application is fun and basically harmless, you may not want your current or a potential employer poking around on your Facebook profile, Twitter feed, personal blog, or photo albums. In 2013, BuzzFeed and CNN collaborated on an article called “ten people who learned social media can get you fired.”106 Examples included “The Bitter Barista” who blogged what he really thought about his customers. Former California Pizza Kitchen server “Timothy DeLaGhetto” tweeted as @Traphik about how little he liked the uniform he was required to wear. He no longer has to wear it since he no longer works for the company.
I once found myself at the very center of a photo mystery that was bedeviling tech journalists. “Who is this geek supermodel?” they were asking as the same woman’s face appeared in both Microsoft ads and promotional materials for archrival LinuxWorld. People were wondering why Microsoft would use a model that was also being used by other tech companies.
“Nobody knows her name,” Robert MacMillan wrote on wired.com. “With immaculately coiffed blond hair, striking black glasses and a perky grin, she’s an idealized image of the geek girl next door.”107 One industrial psychologist said she was beautiful but also somehow approachable, the kind of girl that a guy who has been coding in his parents’ basement for three days could still fantasize about asking out.
I was able to solve the mystery because she was sitting right in front of me in a University of Calgary classroom. Her name is Marla, and, in addition to being a student, she had a job at a local stock photo images firm. “One day,” she told me, “they came around with a camera and took pictures of all of us and we signed model releases.”
Her employer was taken over by Getty Images, which continued to market Marla’s photo. A raft of tech firms decided to use her image in their advertising, often choosing the cheaper “royalty free” option, which meant other firms could use it too.
“It’s become kind of a ‘Where’s Waldo’ type of thing,” Marla told me, “with friends emailing me to tell me about places where they’ve seen my image.” I had a lot of fun breaking this story, and enjoyed her comment that “I love life when it throws bizarre incidents like this my way … I fully recognize that this is my 15 minutes of pathetic fame—so I’m savoring every moment!”108
Revisiting the Marla story a decade later reveals the disturbing way in which photos can persist. The websites for LinuxWorld Expo 2003 and Microsoft’s brochure from that era are long gone. But Getty Images still has her image for sale, and why wouldn’t they? She was a best-seller, even though she told me she never received any royalties for her photo.
For the “royalty free” license fee of less than ten dollars, you too can use Geek Supermodel Marla, at least the way she appeared in 2003, in your next advertising campaign:
Figure 7. Marla in her original form. Getty Images, photo E013748, used under license.
I decided to see if Marla, or at least her hard-working photo, was still on the job. Luckily, we now have tools for that very purpose. One is Google Images search, which allows you to plunk in an image and traverse the Internet, looking for uses of it. This comes in handy for companies like Getty Images to track the use of their photos online.
Doing that turned up 17 hits, including a staffing agency in Lee’s Summit, MO; a web developer in Saratoga, CA; and dentists in both American Fork, UT, and Santa Rosa Beach, FL. She does have lovely teeth but I wonder if the dentists have paid to use her smiling face?
Popping Marla’s iconic photo into TinEye, a Toronto-based competitor to Google Images, produced other uses of her image, from as far away as Poland, Korea, and Japan, mostly related to a Kyocera Zio smartphones ad campaign in 2010. While almost all Microsoft sites have long since dropped Marla, Microsoft New Zealand is still using her smiling face to promote ancient versions of its products.109
We could find even more Marla ads if we used the Wayback Machine, whose mission is to save as much of the Internet’s content as possible for posterity. Just go to www.a
rchive.org, plug in a website, and select a date—you are able to travel through time, browsing cached web pages that have been preserved for eternity.
We used to be able to keep our pictures private. They were in a camera, a trusted photo lab, an album, or a dresser drawer. But even if we are scrupulous about not posting anything online, hackers and even automated technology can spread our photos far and wide. For instance, a Trojan Horse call PixSteal can sneak into your computer and send all the photos it finds to an FTP server waiting somewhere in the world. Coupled with your IP address, and soon, effective facial recognition, the bad guys behind PixSteal are in a perfect position to blackmail you if there is even a single picture on your computer that you would not want made public.
The motivation to find out more and more about you has become a matter of Dollars and Pounds and Euros, Yen, and Renminbi. If Professor Acquisti can de-anonymize people on dating sites from their Facebook profile shots, driven solely by academic curiosity and using publicly available resources, what can somebody do with your personal information when there is real money on the line? A photograph is, after all, an extension of our sense of sight over space and time. But we have other senses too.
Sensor Creep
We have been extending our senses as far as technology will allow since Galileo and his contemporaries turned their optical instruments skyward. Telescopes and binoculars can be used for good (a sailor finding land) and evil (an unwanted voyeur). Sometimes the virtue or vice depends on your point of view. The hunter thinks his binoculars and telescopic gun sight are wonderful tools; the deer in the cross-hairs feels differently.
Everywhere we go, swarms of sensors are watching us. They are in the road, the signs, and the streetlamps. They are in your dishwasher and will soon be in your toothbrush. They are definitely in that “red light camera” that just snapped your car’s picture. You know about that one because you saw a flash and are now awaiting the bad news in the mail. However, most sensors are silent, unlabeled, and often almost invisible. They are talking to each other all the time. Sometimes they let us in on the conversation, sometimes they do not.
Technocreep Page 6