Am I the only one creeped out that Facebook is now guessing, sometimes correctly, if its users are gay?
In the world of Big Data, our private lives are increasingly becoming intermingled with the shadowy, yet public, world of cyberspace.
Whenever we go online we are providing data that can be used to market to us; from Google searches to Facebook likes to eBay purchases, we are inputting data into a series of mathematical models which make incredibly educated guesses about the kinds of people we are.
Facebook creepily offers help to a gay guy thinking of “coming out”
Enter Matt. As BuzzFeed notes, Matt was your typical Facebook user who suddenly found an ad in his news feed for help in coming out. The weird thing was that Matt “did” need help coming out, and understandably he was more than a bit curious as to how Facebook knew.
At first, Matt wondered if Facebook had accessed his text messages, as he had confided in a close friend the previous night that he was gay, and asked for advice. There didn’t seem to be any other reason for Facebook to suggest Clemons’ page; his only two “likes” were for a politician and a local bar, neither of which would suggest anything about his sexual orientation. Moreover, he hadn’t mentioned anything on his wall or over Facebook chat.
After checking his activity log, all he could find were two comments, linked to his Facebook, that he left on a Buzzfeed article covering Rob Portman’s endorsement of marriage equality.
While Facebook insists that it does not access text messages or chats sent through its mobile app, it doesn’t need to directly invade our privacy in order to make startlingly accurate inferences about us.
How they know if you’re gay
Matt’s two comments, along with other data points Matt had provided through his age, gender, browsing habits/history and other trends, led Facebook to estimate that if you lined a hundred people with characteristics identical to Matt’s, a significant number of them will be closeted gays. In short, while Facebook didn’t know Matt was gay and struggling to come out, Matt’s digital profile suggested that he was more likely than your average user to be in that situation.
All the way back in 2009, two MIT grads were able to write a program that was able to discern whether a particular Facebook user was gay based on: A) whether they didn’t specify a sexual orientation in their Facebook profile; and B) whether they had an inordinate number of openly-gay Facebook friends. There are a lot of different ways to guess whether someone’s gay.
To put this in a real-world context, with some employers now asking for access to job-seekers’ Facebook log-in information, do you really want every future employer to know that you’re gay, which they’ll suspect as soon as they start seeing all the gay ads popping up once they’ve logged in as you. For some, that may not be an issue. For others, it most certainly would be. But regardless, it should be your decision to release that information, and to whom – no one else’s.
And how you vote
This form of marketing is here to stay, and is becoming more ubiquitous over time. From Target not only knowing that you’re pregnant, but being able to predict your due date within an impressively small window, to the Obama campaign knowing who you were going to vote for even before you did, targeted modeling of individual behavior is creeping into an increasingly large share of our lives. Of course, it didn’t start with the Obama campaign — John Aravosis wrote about this predictive-voting phenomenon for the Economist back in 2000.
On its face, why shouldn’t companies and political organizations use this method of targeting shoppers and voters? It’s effective, relies on ostensibly public information, and increases our ability to satisfy our needs and wants. After all, Matt did need help coming out, and that woman visiting the Target Web site really was pregnant, right?
But even if these models rely on seemingly-benign information that we make public, and make our lives more efficient (analytics are used in everything from suggesting book purchases to city planning), it still feels uncomfortable when a non-human entity knows so much about us. Though I suspect gays and lesbians would hardly be comforted if it were a human being secretly cataloguing their sexual orientation either.
Logically, individual-level marketing is no different than a clerk at our local department store remembering the kind of shoes we like, or a bartender who always knows our favorite drink, so why does it feel like an invasion of privacy when a computer does it?
Why you might be feeling creeped out by all of this
The answer may lie in, of all places, “road rage.” In 1995, researchers at AAA reported that, while the roads we drive on are public, we consider the area in and around our cars to be our personal space. Drivers who encroach on such space are seen as invading our territory:
“As individuals we have a personal space, or territory, which evolved essentially as a defense mechanism—anyone who invades this territory is potentially an aggressor…The car is an extension of this territory. Indeed, the territory extends for some distance beyond the vehicle…providing room for the defender to prepare to fend off or avoid the attack. If a vehicle threatens this territory by cutting in, for example, the driver will probably carry out a defensive maneuver. This may be backed up by an attempt to re-establish territory…flashing headlights or a blast of the horn are, perhaps, most commonly used for this purpose. However, this may not always succeed in communicating the full depth of our feelings. As it is usually difficult to talk or even shout to the offending driver, other non-verbal communication (offensive gesticulations) may be employed…In some circumstances, the defending driver may wish to go one step further and assert his dominance….This is comparable to the manner in which a defending animal will chase an attacker out of its territory.”
While we obviously don’t have the capacity to honk a horn or get in a fistfight with Facebook, the relationship between personal and public space is similar online as it is on the road.
While our activity online is often very much public, it happens within our own personal space, alone on our computers. And we have an expectation of privacy. I feel like my Google search and Web comment histories are, well, mine. The idea that they can be used to sell me products or provide me with life-coaching resources feels like an invasion of my personal space. Even if, as I mentioned above, the same documentation of our tastes and habits is happening when the waiter at our local diner asks us if we want “The Usual,” the fact that our online activity is an individual pursuit makes it feel like the proverbial online waiter shouldn’t know what our “Usual” is. Much like on the road, we seek to maintain our bubbles of personal space in an undoubtedly public online world.
The Internet changes things
In the old days, the bartender knew your favorite drink, the waitress your favorite meal, the barista your coffee, your travel agent when and where you liked to fly, and the local pharmacist what kind of prescription drugs you took (and thus, your ailments) and possibly what kind of birth control you used, and how often you used it. In the case of the postal service, sure you gave them confidential correspondence, but you generally knew that they’d never be peeking inside. And in the case of the phone company, if you had a little sexy talk with your significant other, no one was the wiser because you had a reasonable expectation your amorous dialogue wasn’t being stored for posterity.
But with Big Data, a lot of those once-private activities are happening in the same locale, via the same venue. The same person, as it were – your computer – is handling your food, clothing, travel, prescriptions, and even phone conversations and dating. It’s as if the little old lady at the corner 7-11 didn’t just know what your favorite Slurpee was, she now has your entire medical and sexual history, and she’s perusing your phone calls and opening your letters to make sure you’re not a pervert, or even simply sharing music or movies with your friends. By aggregating seemingly-public content, companies today are able to invade our privacy in ways formerly unimaginable.
In terms of Big Data, it’s almost impossible to draw lines that divide acceptable and unacceptable business practices. As the BuzzFeed article noted, “[The use of analytics is] normal, in that weird way where saying the same thing — ‘Facebook knows I’m gay,’ or ‘Target thinks I’m pregnant and due in August’ — can sound like either total paranoia or smart business, depending on how you say it.” The same models used to accurately predict Matt’s sexuality are used to more efficiently link us to the goods, services and information we use on a daily basis. Moreover, while, for Matt, the result came off as an uncomfortable breach of privacy, it only takes a handful of public data points to tell the world a lot about yourself.
Consider that fact before you make your next click.