NYS Exposed: DMV using facial recognition tech to crack down on fraud

August 24, 2017 11:43 PM

ROCHESTER—21,000 people have been busted in New York State in the past few years trying to use stolen or bogus information to get a driver’s license. The Department of Motor Vehicles has been cracking down on the fraud by using updated facial recognition technology.

Every time your photo is taken at the DMV, it is now run through the software. The photograph is compared not only to the last photo that was taken of you but to all the photos that have been taken at DMV offices across the state.

The facial recognition technology uses 128 measurements to compare; different hair styles, glasses and other features that change over time do not prevent the technology from making matches.

“About half of the people we identified were driving the roads while they were suspended or revoked under another name,” says Owen McShane, the Director for Investigations for NYS DMV.

So, the technology has arguably made the roads safer, it’s also stopping fraud: ”A lot of the individuals are using a second identity to commit benefit fraud, hide the fact that they are wanted under another name or just hide from somebody,” McShane adds.

One scammer, denied a license here in New York, flew to Florida and got one there. He then leased a car, bought furniture and drained $50,000 from someone’s account. "When we showed up at his apartment to arrest him, Raymour and Flannigan showed up to deliver furniture he'd purchased with a stolen identity as well, so he had a really rough day, the day we arrested him,” recalls McShane.

While the technology has put criminals behind bars, it does have its critics. The New York Civil Liberties Union is not completely on board. In a statement to News10NBC a spokeswoman says, “Facial recognition technology has very low accuracy rates, especially for young people, people of color and women. Given those severe flaws, the public is entitled to know exactly how the DMV is using facial recognition and the department must place strict limits on how the results are used.”

Jiebo Lue, a computer science professor at the University of Rochester, says while accuracy issues may have been a concern a few years ago, the technology has made huge strides since 2015. "That's the first time when reported accuracy was above 99.5% when you do a one to one match and that number basically exceeds human performance,” he says.

While state and federal agencies, banks and retailers have been slowly investing in the technology here in the United States, in some countries like China, it’s already part of citizens everyday lives. "In public restrooms, they have paper towels…and some people take more than what they need so now they have face recognition technology to say within 10 minutes the same person can only take this amount,” Luo says.

While it’s unlikely the technology will ever be used to that extreme here, more companies and agencies are starting to implement it. Recently, the Department of Homeland Security asked for bids from contractors who could use facial recognition on drivers coming into and out of New York City through the tunnels and bridges. It pulled back the request though after being told by industry experts that speed, lighting and window glare would significantly hinder the accuracy of the readings.