Last year we editorialized about the use of facial recognition software by government ("In Favor of Access to Facial Recognition Technology for Law Enforcement"). We pointed out that this technology would greatly increase the power of the state to track and record the movements of individuals who are not suspected of any crime but may otherwise be of interest as, for example, political critics, as it is being used abroad. We also pointed out that its validity and probative value when used in law enforcement would be tested by the same process and the same standard that has been applied over the years to fingerprints, ballistics, drug sniffing dogs, DNA evidence and other technical forms of proof, including technologies like voiceprints and the polygraph that have been rejected as unreliable.

That day of reckoning has now arrived in New Jersey because the use of facial recognition software has led to the false arrest of an indisputably innocent man. In January 2019, the Woodbridge police confronted an alleged shoplifter. After presenting them with a Tennessee driver's license that turned out to be a fake, he escaped in his rental car. The police used facial recognition software to match the photo on the license with a state database of photographs, and it turned up the state-issued photo id of Nijeer Parks, an African-American New Jersey resident. Parks, who testified that he does not drive and has never been in Woodbridge, was eventually able to establish an ironclad alibi – he was sending a money order from a drug store 30 miles away at the time of the crime. The charges were dismissed, but not before Parks had spent 10 days in jail and spent $5,000 in legal expenses to defend himself. He was fortunate in his alibi. Without it, he might well have been convicted on the bases of a false identification supported by a spurious claim of scientific conclusiveness.