Facial recognition tech nabs ID fraudsters
- By Matt Leonard
- Aug 31, 2016
A man with a suspended commercial driver’s license attempted to get a new one using a stolen identity, but he was stopped. About two dozen people used fake information to get a second Social Security numbers and try to get new licenses, but they were stopped. Five individuals tried to take over someone else’s account at the Department of Motor Vehicles, but they were stopped too.
All of these people were foiled because of New York DM’s use of facial recognition software. The technology was first piloted in 2008, and as the system improves, it is steadily detecting more fraud attempts.
The visit to the DMV doesn’t change for a visitor. The same camera takes a picture of the driver’s license applicant. Then an algorithm developed by Morpho Identity Systems takes facial measurements and uses that data to compare the new photo to an existing DMV picture – and to the agency’s whole database of images -- to ensure drivers are who they say they are. The algorithm flags images associated with more than one account.
During the pilot phase of the program, the algorithm was able to measure 32 points on a person’s face. That number improved to 64 when the program went live in 2010. Now it tracks 128 points.
The more points the system can recognize, the better job it will do in finding a fraudster, according to Owen McShane, an investigations director at the New York DMV. In the past, he said, the system struggled to match images of people whose hair and glasses had changed, but now those issues have been minimized.
The latest upgrade has led to 100 arrests and 900 open cases since it was launched in January, according to state officials. In six cases people have tried to renew another person’s driver’s license, McShane said.
Each night the algorithm compares new DMV photos to the database and provides a list of potential matches to workers in the morning. This list is double-checked by humans because old photos and changes in physical appearance can sometimes lead to false positives, McShane said. New licenses are not issued until a photo clears the DMV database.
This database is not shared with other law enforcement agencies, but if the DMV finds a duplicate license, it makes sure it’s not interfering with a separate investigation before proceeding, McShane said.
"The results from our use of this enhanced technology are proof positive that its use is vital in making our roads safer and holding fraudsters accountable,” New York Gov. Andrew M. Cuomo said in a statement.
New York is far from the only state to have implemented facial recognition technology in the effort to thwart identity fraudsters. There were at least 39 states using the technology in 2015, according to Stateline, including New Jersey, North Carolina, Vermont and Iowa.
Additionally, New York is sharing its database with the New Jersey DMV, which resulted in finding 62 people who held commercial driver’s licenses in both states. The states initiated this pilot in response to a number of serious crashes involving commercial drivers who possessed multiple licenses issued by different states.
Matt Leonard is a former reporter for GCN.