The state's aggregated investigative reporting service allows secure and real-time access to fraud tip-off information provided by 23 law enforcement, state and federal agencies.
If it wasn't for facial recognition technology matching Rickie Storie's three driver's licenses to his three identities, the New Jersey truck driver might not have been caught and convicted after his 64 suspensions and six DUI convictions.
"This is someone who clearly should not have been behind the wheel of a car, much less a tanker-trailer, much less a tanker-trailer with hazardous material endorsements," said Raymond Martinez, chairman and chief administrator of the New Jersey Motor Vehicle Commission.
The state attorney general's office charged Storie and 106 others with fraud, thanks to the commission's facial recognition project -- dubbed Operation Facial Scrub or FS-12 -- that matched pictures from New Jersey's photo record database to uncover drivers who had multiple licenses under various names, much like Storie. County prosecutors' offices also found fraud and have filed additional charges.
The commission and the division of criminal justice partnered to conduct FS-12, which identified 1.8 million total photo matches, including 2,300 fraud cases and nearly 1,000 potential criminal cases.
The technology uses algorithms from numerical points on the human face to find photo matches. Facial recognition technology has become faster and more accurate, and is being used successfully in many state motor vehicle applications to identity fraud cases, ensuring that each person has one identity.
While the operation's "scrub" ended early this year, the fraud-fighting is ongoing -- daily scrubs are helping the commission to continue to catch dishonest drivers, while a secure website is sharing possible fraudsters with other state and federal agencies to extend the project's reach.
"It's an interesting project," Martinez said. "And the benefits are not just for motor vehicles now -- the benefits are for all of our state and federal partners as well."
Reasons to obtain a fraudulent identification range from fleeing from past misdeeds like drunken-driving convictions to pursuing ongoing criminal activity such as cashing fraudulent government checks. Most of the defendants nabbed by the FS-12 operation acquired new fraudulent licenses after theirs had been suspended.
However, fraud can be committed against any number of government agencies because a state-issued driver's license has become the most relied upon identity document in the country, said Thomas Flarity, director of security, investigations and internal audit for the commission. He noted the 9/11 Commission's findings that all jurisdictions should be enhancing the reliability of driver's licenses.
That's why in 2011, New Jersey began planning for the massive "scrub" of its legacy database, then containing 19 million photos of more than 7 million drivers, plus those who had left the state or were no longer living. The database now has 23 million.
The New Jersey Motor Vehicles Commission is in talks with New York Department of Motor Vehicles to expand on both states' facial recognition programs, testing for fraud that crosses borders. The project, which is still in the works, will focus on commercial drivers who have been known to create false identities in different states to stay driving and employed with infractions of DUIs, or as a fail-safe before something happens.
The interjurisdictional undertaking is the first of its kind addressing driver's license fraud, said Raymond Martinez, chairman and chief administrator of the commission. And it's needed -- there is no system for tracking driver's license fraud between or among states; the National Driver Registry only works for someone who keeps their same identity, said Thomas Flarity, director of security, investigations and internal audit for the commission.
"The only thing you can't get past," he said, "is that your face is the same in both instances."
The state first glimpsed its fraud problem -- which plagues all states, Martinez said -- a few years back when its program to match Social Security numbers to identity revealed many cases of it. And, nearly half of all states at the time were already using facial recognition in the motor vehicle arena, he said.
In addition to its applicability, it was also clear from surveying other states that significant work would result from FS-12, Martinez said. Although facial recognition is sound technology, enough people look alike that there are false positives.
Retired law enforcement officers hired for the forensics task went through the legacy photo matches, separating the lookalikes and clerical errors from the genuine cases of fraud. Although not as "sexy" as the criminal cases, Martinez said that staff uncovered around 6,000 cases of clerical errors helping the commission to save its customers time and headaches.
The FS-12 undertaking began in early 2012 and took two years. "We were operating on all cylinders here and responding to a lot of incidents of fraud," Flarity said.
While the cases of driving fraud were explicit, in some instances, the commission found it difficult to pinpoint the type or extent of fraud just by finding the multiple identities; sometimes, the driver would have a perfect record on both licenses, Martinez said.
"There's a possibility that they had ongoing fraud of another nature that we weren't privy to -- they hadn't been caught," he said.
Four months after the FS-12 launch in February 2012, to help unearth more fraud, New Jersey revealed a secure website called AIRS, or aggregated investigative reporting service.
The commission created the website with the help of the office of information technology as a way to kill two birds with one stone with its FS-12 operation, systematically busting fraud both on and off the roadways.
The idea first came from other states identifying fraud in driver's licenses. The states would stumble on this ability to fight fraud in the wider government realm, Flarity said. This spurred New Jersey to ask: "How could we strategically identify that fraud rather than stumble on it?" he said. "We believed that if we attacked this strategically and affirmatively on the outset, we were going to get the best bang for our buck."
A website was then proposed and cemented during a large-scale seminar in December 2011, put on by the commission. Attendees included New York and North Carolina representatives, New Jersey law enforcement, and state and federal agencies, 23 of which signed up for the ability to use the site. Federal participants include the Joint Terrorism Task Force, the Department of State-Passport Security and the Social Security Administration.
By signing up for the website, which allows secure and real-time access to the fraud tip-off information, those agencies agreed to certain parameters like treating the information as confidential and following up to report fraud so that investigations could be coordinated.
"The whole point of the AIRS site was to encourage collaboration between the governmental entities that would have an interest in an identity fraud case," Martinez said. "And secondly to de-conflict investigations and prosecutions, and put them together to the extent possible for the efficiency of all the agencies involved."
Flarity said that from the first batch of 250 cases posted on the AIRS website, the Social Security Administration opened 70 case files, one of which was for a woman with two identities who had for 20 years been collecting disability benefits while working. Many other agencies had similar success stories, especially coming out of the gate, he said.
In the first year, the commission shared the list of people suspected of fraud, dozens of names long, on the AIRS website about once every two weeks. Now, because the facial recognition technology is recognizing only new fraud, about 10 people are posted to AIRS in two weeks' time. If a fraud is identified, the agency is directed to e-mail the site's central drop box, watched by the Division of Criminal Justice.
"We have people there to collate and coordinate the mutual interest of everybody so that one targeted approach to that individual could be made on behalf of all potential victims, including motor vehicles," Flarity said.
Martinez added that some entities, like the Department of Labor and Workforce Development, which runs the state's unemployment benefit program, have created comprehensive systems of their own to vet potential fraudsters and report back. And for each person, law enforcement does its own due diligence to assess whether the case is a criminal matter, and either the attorney general's office or a local prosecutor addresses it.
Though the state's photo database is now clear of fraud, Martinez said people continue to create of new fake identities. That's why the state has continued to use the facial recognition software to run a daily scrub on any new photo, and to post the fraud suspects to the AIRS website.
While the AIRS architecture is sound and government agencies continue to benefit from the site as many report back, Flarity said that some agencies have tapered off in their reporting or don't report at all. He said he suspects it's not because of lack of fraud.
"When some other agency is bringing in proof to your doorstep that you might be infiltrated or there might be some bad news, not everybody is over-the-top enthusiastic to embrace that," Flarity said.
Even so, Martinez said: "We feel we have the responsibility to share that information."
The commission plans on reconvening stakeholder entities, like it did at the start, to show how the site has helped fight fraud and to retrain them on how to use AIRS to their advantage.
"I think that's really the next step," Martinez said, "to really let it take root not just for motor vehicles and the law enforcement side but also for benefit-conferring agencies."