By Jim Spencer, Star Tribune (Minneapolis) –
WASHINGTON — Imagine arriving at a hotel to be greeted by name, because a computer has analyzed your appearance as you approached the front door.
Or a salesman who IDs you and uses a psychological profile to nudge you to pay more for a car.
The day is coming when businesses, and others, will have those kinds of capabilities, says Alessandro Acquisti, a professor at Carnegie Mellon University who studies the positive and negative implications of facial recognition technology.
“Your phone — or in some years, your glasses, and in a few more, your contact lenses — will tell you the name of that person at the party whose name you always forget,” Acquisti said. “Or it will tell the stalker in the bar the address where you live.”
However it is used, facial recognition is becoming a big business, with the potential to move far beyond such early applications as picking out Facebook friends in photos or helping cops nab crooks and terrorists.
In 2010, Maplewood, Minn.-based 3M Co. paid nearly $1 billion for the California company Cogent, which develops a variety of identification systems, including iris and facial recognition technology. Another company, MorphoTrust USA, has a 150-person biometric facility in Bloomington, Minn.
“The next step in applications will be face-in-the-crowd — identifying people at long distance,” 3M Cogent marketing director Teresa Wu said. “That depends on the resolution of the cameras.”
In 1990s pop culture trivia, Cheers was the bar where “everybody knows your name.” In 21st-century America, with its social media, photo sharing and myriad cameras, you increasingly need only to appear in public to be recognized, even if you don’t want to be.
Researchers say facial recognition technology is well on its way to supplanting fingerprints as the biometric identifier of choice. Eye and facial recognition technology are predicted to account for nearly $4 billion of the biometric industry’s $11 billion in annual revenues by 2017. That same year, the number of cellphones and tablets equipped with facial recognition features is expected to swell to 665 million.
With recognition rates of 99.7 percent for well-lit, frontal photos, research and development has moved to identifying people without their knowledge — whether in crowds or shadows or not looking directly at the camera.
The Orwellian potential of facial recognition has attracted the attention of lawmakers. When U.S. Sen. Al Franken, D-Minn., looked at an explanation of how the FBI plans to use facial recognition technology to identify criminal suspects, he was surprised to find photos of political rallies.
The FBI says it has strict rules about how it uses facial recognition. It can only match photos of strangers against a national database of criminal mug shots. If the photos match the database, the FBI can only suggest a lead for local, state and federal investigators to follow. It cannot offer a positive identification.
Still, some see potential for abuse.
“The FBI shouldn’t be in the business of monitoring demonstrations unless it has a cause, a tip, a reason,” said Gregory Nojeim, senior counsel at the Center for Democracy and Technology. “Speech is easily chilled by that.”
As chairman of the Senate Subcommittee on Privacy, Technology and the Law, Franken finds himself at the center of a larger national debate that increasingly pits information and technology against privacy and free speech.
On one side, you have people like Calhoun County, Ala., Sheriff Larry Amerson, who recently told Franken’s subcommittee that “some advanced facial recognition in use today … is as accurate as fingerprints, but results are obtained in seconds, not hours.”
On the other, you have people like Acquisti, who makes sure people understand how good technology can be put to bad uses. With fellow researchers Ralph Gross and Fred Stutzman, Acquisti proved you could take photos of anonymous students and use the Internet and over-the-counter facial recognition software to identify 30 percent of them before they finished filling out a three-page survey.
The researchers designed mathematical algorithms to extract sensitive personal data associated with the names of those identified. They also tested a cellphone face-recognition app that placed a name to a face in real time.
Meanwhile, Facebook uses facial recognition technology to support its popular tag suggestions feature, building a library of so-called “face prints” that privacy advocates say approaches a billion images and is second only to law enforcement.
Franken voiced concerns about how the social network might use those face prints besides as a tagging tool.
At a recent hearing, Franken asked Facebook’s manager of privacy and public policy, Rob Sherman, if the company would rule out sharing its face-print library with a third party. Sherman would not rule it out. A newly finalized settlement between Facebook and the Federal Trade Commission may or may not require Facebook to obtain permission before doing so.