There Will Be No Turning Back on Facial Recognition
It’s not perfect yet, but it’s already changing the world
By Lane Brown
On Friday, August 16, at around 7 a.m., a pair of suspicious appliances was found on a subway platform at the Fulton Street station in lower Manhattan and, an hour later, a third near a garbage can on West 16th Street. Initially, police thought they might be improvised bombs, like the shrapnel-filled pressure cookers that blew up at the 2013 Boston Marathon and in Chelsea in 2016, but upon inspection they turned out to be harmless empty rice cookers, probably meant to scare but not explode. Trains were delayed during the morning commute, but since that happens often enough without any terrorist help at all, the scariest thing about this episode may have been the way the alleged perpetrator was caught.
Minutes after the discovery, the NYPD pulled images of a man leaving the devices from subway surveillance cameras and gave them to its Facial Identification Section (FIS), which ran them through software that automatically compared his face to millions of mug shots in the police department’s database. The program spit back hundreds of potential matches in which officers quickly spotted their person of interest: Larry Griffin II, a homeless 26-year-old the NYPD had arrested in March with drug…
