Walking down the street, a stranger quickly scans your face from behind a seemingly nondescript pair of glasses -- but hidden in the frames is a tiny camera.

Using embedded wi-fi technology, the person identifies you not just by name, but by place of birth, credit score and Social Security number, in a matter of seconds.

It sounds like science fiction, acknowledged Alessandro Acquisti, an associate professor of information technology and public policy at Carnegie Mellon University. But according to new research Mr. Acquisti and his colleagues presented at the annual Black Hat tech security conference in Las Vegas Thursday, it's all too real.

Using off-the-shelf, facial recognition software by local firm Pittsburgh Pattern Recognition and a bank of Facebook profile pictures, Mr. Acquisti and CMU researchers Ralph Gross and Fred Stutzman were able to discern the names of just over 31 percent of 93 CMU students who allowed the team to take a snapshot of them.

Advertise with us

For about 27 percent of the students, Mr. Acquisti's team was able to use a research protocol they designed two years ago to mine online stores of data and come up with more information about the students -- date and place of birth and, eventually, fragments of the unsuspecting students' Social Security numbers.

"We asked subjects how they felt about that scenario, and, almost universally, they didn't like the scenario. Some freaked out," Mr. Acquisti said with a laugh. "But all of our subjects put the information online that made that identification possible."

When coupled with facial recognition software -- the same technology Facebook has been using since June to allow users to more easily identify and tag friends in photos -- internet users' urge to overshare spells trouble for privacy hawks.

But while the list of characteristics Mr. Acquisti said are identifiable using just a photograph is long and includes traits such as religious and political affiliation and sexual orientation, he said there is no cause for alarm.

"We are not doing this to scare people at all," he said. "We are doing this to raise awareness about what in the future, which is no longer so distant, will be able to be done by almost anyone."

Advertise with us

Mr. Acquisti said his study also demonstrates the weakening boundary between online and offline information: Something as real and tangible as a face can unlock treasure troves of digital information that Web users are willingly relinquishing on a daily basis, he said.

Facial recognition software will grow more accurate as developers refine the technology, expanding the effect Mr. Acquisti identified in his study.

While Google has mulled use of the technology for enhanced searches, Google spokesman Eitan Bencuya said the company has resolved not to include facial recognition technology by Pittsburgh Pattern Recognition, which the tech giant acquired in late July, in applications or product features without "strong privacy protections in place."

While Mr. Acquisti said he and his team could continue to study the measurable effects of facial recognition software upon online and offline privacy, he plans to turn his attentions to the way the compulsion to share information is winning out over the urge to protect it.

"This paradox where we want privacy but engage in activities that endanger our privacy is something I remain interested in investigating," he said.