August 2, 2011
Facebook Becoming Worldwide Yearbook Of Identification
According to a new study, one in three people can be identified on Facebook by using just a photograph.
Researchers took a picture of 93 volunteers to use this method to track them down on the social networking site by using publicly available software.
The researchers were able to predict the first five digits of their social security number 27 percent of the time by using the volunteers Facebook profile.
The study used the Pittsburgh Pattern Recognition (PittPatt) facial recognition software.
The Carnegie Mellon University researchers took pictures of the 93 students and compared them with 261,262 publicly available photos downloaded from Facebook profiles.
The system turned up 10 possible matches, and the correct photo was in the top three over 30 percent of the time.
Professor Alessandro Acquisti, the lead researcher, said in a press release that technology had now evolved so that there was a "democratization of surveillance."
The team said Facebook is becoming the world's default identification service.
The only available pictures in the past have been school photographs or mugshots of criminals.
Now, ID thieves could potentially find out everything they need to know by looking through the world's most popular social network.
Acquisti said the study "suggests that the identity of about one-third of subjects walking by the campus building may be inferred in a few seconds combining social-network data, cloud computing and an inexpensive web cam."
Acquisti was able to identify the first five digits of the test subjects' social security numbers with just four attempts, which could be a potential money-making opportunity for ID thieves.
"The chain of inferences comes from one single piece of anonymous information - somebody's face," he said in a statement.
Paul Ohm, a law professor at University of Colorado Law School, who has read the paper, said the research showed how easy it was to "re-identify" people.
"This paper really establishes that re-identification is much easier than experts think it's going to be," he told The Wall Street Journal.
According to the National Institutes of Standards and Technology the proportion of photographs wrongly identified by facial recognition programs has dropped from 79 percent in 1993 to just 0.29 percent last year.
On the Net: