March 5, 2013 at 5:55 PM ET
New York state officials said Tuesday the DMV's use of facial recognition software for driver's licenses since 2010 has resulted in more than 2,500 arrests of those trying to steal someone else's identity or trying to get a second license.
"Through this program, we are successfully taking dangerous drivers off our roads, helping to track down criminals, and protecting taxpayer dollars — sending a clear message that New York State does not tolerate identity fraud and those who try will be caught," Gov. Andrew Cuomo said in a statement.
New York's DMV has investigated 13,000 possible cases of identity fraud using facial recognition, with 2,500 arrests resulting and "more than 5,000 individuals facing administrative action," according to the governor's office.
The DMV started using its facial recognition software to help identify persons who may have had their license suspended, but are trying to get another license, or who are trying to steal another person's identity. The system uses software algorithms to compare new photos taken at the DMV with existing photos in the agency's database. If there's an inexplicable match, then an investigation by "trained staff"ensues.
"This review includes new photos taken each day at the DMV, as well as approximately 20 million photographs already in DMV’s database," the state stays.
The sci-fi-like enhancement to the licensing process is not is not hailed by all. Watchdogs fear the license photo database being shared with law enforcement and other agencies.
"One potential problem, from a privacy standpoint, is the sharing of this facial recognition database with other governmental and non-governmental entities," Ginger McCall, director of the Electronic Privacy Information Center's Open Government Project, told NBC News.
"The Federal Bureau of Investigation and Department of Homeland Security have already begun testing crowd-scanning facial recognition technology in stadium venues," she said.
"Pairing crowd-scanning technology with DMV facial recognition databases would allow for the wholesale surveillance of innocent Americans — potentially tracking and databasing people's location and protected speech and associational behavior."
Among the DMV statistics, as analyzed by the University at Albany’s Institute for Traffic Safety Management and Research, for people who had two or more licenses:
Approximately 50 percent of the subjects identified through facial recognition had one valid NYS license while having a second record that was suspended or revoked.
Approximately 20 percent of the subjects identified through facial recognition were suspended or revoked under every known record.
Approximately 30 percent of the subjects identified through facial recognition had multiple valid licenses.
Felony arrests that resulted from facial recognition included more than 100 people who had "active felony warrants" under one driver's license name, but set up a new state ID under an alternate name. "One of these subjects was a fugitive for 17 years after robbing a bank in Nassau County in 1993," the state noted.
Another case involved a man who was working as a school bus driver under one name, but who had "multiple open suspensions for unpaid tickets, as well as narcotics convictions and suspensions of his license privileges for narcotics transactions. The subject was arrested under multiple felony charges and is no longer operating a school bus."
New York is not the only state using facial recognition for driver's licenses; several are now. And no recognition system is foolproof. The ACLU warns about that concern in a recent blog where it noted that in 2011, there were at least 1,000 false reads by Massachusetts Registry of Motor Vehicles, "causing pretty substantial inconvenience to people accused of fraud by an imperfect computer program, and mucking up the system for everyone else."