Humans may claim to enjoy surprises, but everything about our behavior suggests the opposite. We'd rather know and we'd prefer to know ahead of time. The instinct is visceral, borne of experience gleaned on the open savannahs in ancient Africa where surprises were rarely good and survival might well depend on the difference between sensing subtle changes in the world around us.
Flashing forward to the digital age, we continue to carry this genetic imprint. From capital markets to web-based commerce, we prefer to know as much as possible in order to lower our costs of inventory and supply chain management as well as to increased the probability that resources invested in a product or service will produce the optimal purchase decision.
So it should come as no surprise that the algorithms behind Facebook's 'Like' Button reveal much broader and deeper types of personal information. Statistical analysis is designed to suss out correlation and causality. When aggregated across a billion-strong user base, certain basic truths become apparent: those who like one thing are more likely to like another or be of a certain age, religion, nationality, sexual preference, color or income cohort.
Some of this will increase access to deals or products or services the research shows they may be inclined to find attractive. This, in turn, enhances Facebook's ability to deliver customers and profits to advertisers and other commercial partners. Given the company's travails in convincing investors that its growth is not just in grandmas posting photos, this sort of digitally driven business could be significant. The privacy concerns that go with the knowledge may be troubling, but there are few who appear to believe there are any secrets left once they connect to the web. The question is whether secrets revealed, however inadvertently, raise broader societal concerns about personal space. The evidence to date is that such an outcome is unlikely anytime soon. But as we continue to test such boundaries we must be prepared to find that at some point we will cross a line too far. Any maybe such algorithms will help determine where that is. JL
The Financial Times reports:
Facebook’s ‘like’ button can reveal more than you realize, a new study has showed. By liking posts and links, you may be revealing personal secrets, like your sexuality or religious and political views.
The findings have raised privacy concerns.
A study by the National Academy of Sciences examined 58,000 Facebook users in the US, who volunteered their likes, demographic profiles and psychometric test results. Researchers managed to draw “surprisingly accurate” findings about a given user’s race, IQ, sexuality, substance use, personality and political views by analyzing the topics and items they ‘liked,’ even if they set strong privacy settings for their page.
The study’s authors developed an algorithm that uses Facebook ‘likes’ to create personality profiles, potentially revealing a user's personality. Anyone with training in data analysis could be able to derive such information, even if users had not explicitly shared it, they explained.
As a result, researchers were able to predict whether men were homosexual with 88 percent accuracy by their ‘like’ clicks on sites related to gay marriage or same-sex relations. Preferences of music and TV shows, for example, were also more revealing than users may have imagined: Men who liked the musical TV show ‘Glee’ were more likely to be gay, the study showed.
In 82 percent of cases, Christians and Muslims were correctly identified among the volunteer profiles. And the study was not only predictive of sexuality or religion, but also a user’s IQ.
Those with higher IQs tended to more frequently like ‘The Colbert Report’ TV show, for example, or films like ‘The Godfather’ and ‘To Kill a Mockingbird.’ Those with lower IQs liked Harley Davidsons and Bret Michaels of the rock band Poison.
Similarly accurate predictions were also made about users’ political views, and could even predict potential voting patterns during elections, mapped to users’ relationship status, their number of Facebook friends and a half-dozen other personality traits.
“The important point is that, on one hand, it is good that people’s behavior is predictable because it means Facebook can suggest very good stories on your news feed,” South China Morning Post quoted Michael Kosinsky, one of the academics behind the study as saying.
While the study could be seen as a new step towards creating more personalized content, it also highlighted the potential threats to privacy posed by Facebook. “What is shocking is that you can use the same data to predict your political views or your sexual orientation. This is something most people don’t realize you can do,” Kosinsky warned.
Similar profiles could be created using other digital data, including Web searches, emails and mobile phone activity, according to a co-author of the study: ``Your likes may be saying more about you than you realize,'' David Sitwell, a Cambridge University researcher said, according to AFP. "But you don't realize that years later all those likes are building up against you."
Facebook declined to comment on the study.
0 comments:
Post a Comment