Be Careful With That Facebook Curly Fries Like
They know everything about us, or so it seems. Facebook, Twitter, Google, or almost any website you use is watching you and your every move. But how much do they really know and how accurate are their predictions about our behavior? Well, if one report is anything to go by, there is a high degree of accuracy, but some of the predictors are, on the face of it, bizarre to say the least. It seems, for instance, that among the top predictors for high intelligence is a liking of curly fries. How can that be? The article lists 11 likes most predictive of high IQ. These are “liking” The Godfather, Mozart, thunderstorms, The Colbert Report, Morgan Freeman’s voice, The Daily Show, The Lord of the Rings, To Kill a Mockingbird, science, and of course those curly fries.
The top predictors of low IQ were posting likes for Jason Aldean, Tyler Perry, Sephora, Chiq, Bret Michaels, Clark Griswold, Bebe, I Love Being a Mom, Harley Davidson, and Lady Antebellum.
Other attributes studied by the report included age. Older ages were predicted by liking Coffee Party Movement, Dr. Mehmet Oz, Small Business Saturday, and Fly the American Flag, while the younger age groups were predicted by Walt Disney Records, I Hate My ID Photo, Body by Milk, and Dude Wait What.
You might think that some of those selections would give an indication of likely intelligence, but most of them, I’m sure you would agree, seem to have little or no connection with intelligence and the curly fries like is a good example. There appears to be no obvious correlation here between the content of the like and the trait it is predicting.
To understand what is going on here, we need to look at what sociologists describe as â€śhomophily,â€ť which is the well-observed phenomenon that people tend to like other people who are like themselves. Smart people, therefore, hang around other smart people. It is not hard, then, to see how a liking by one individual for a particular book, band, or TV program can be passed on to another and, in a mirror image of, say, a YouTube video going viral, spread like crazy. This is explained in a fascinating speech by Jennifer Golbeck. So, she claims, just one smart guy liking curly fries can go viral among his oh-so-smart friends. In the world of Internet personality predictions based on the “like” system, the content of the like can be irrelevant to the trait it predicts.
In her talk, Professor Golbeck shows how Internet companies gather information in vast quantities. She points out that half of the world’s Internet users interact with Facebook, which, like other companies, uses the data to drive sophisticated but secret modelling processes to turn that data into dollars. In truth, the Internet giants don’t just base predictions on simple one-off likes, but use multiple individual online activities to build up a picture that will include assumptions about our gender, age, sexual orientation, race, religion, intelligence, political leanings, even the strength or otherwise of our relationships with others.
One example Golbeck gives is that of a report in Forbes magazine of a 15-year-old girl receiving a flyer from Target promoting baby bottles and diapers. Two weeks later she told her parents she was pregnant. Target had correctly put together information from the girl’s online activity to predict she was indeed with child.
It’s scary stuff and it’s not hard to see how abuse of data can get more problematic than it already is, but Golbeck’s speech offers up some interesting ideas on how the average web user may be able to claw back some control from the hidden modellers and the fortune telling algorithms. Quite rightly she thinks that waiting for the politicians to change the law is not the solution. It is not likely to happen any time soon. In the same way, expecting or waiting for Internet companies themselves to change their policies is a waste of time. As she points out, many people regard Facebook users not as customers, but as product — data equals dollars. Why would the shareholders of Facebook want to restrict their earnings potential in a hurry? Golbeck proposes that some form of elective process, such as the user deciding who can or cannot access their data, is one way to go. Another is an increased use of encryption that denies the data predators access to your online lives. The technology is there to do this already. What is lacking is the will to drive it through.
Image Credit: Thinkstock