Police Arrest Impossible Suspect After Facial Recognition Blunder
A 6-foot-2-inch-tall man was wrongfully arrested after New York Police’s facial recognition technology linked him to a suspect who was significantly shorter and lighter than he was.
Civil rights and privacy groups are calling for an investigation into the New York Police Department’s (NYPD) use of the system following the false identification and wrongful arrest of Trevis Williams.
According to The New York Times, on February 10, a woman reported that a delivery man had exposed himself in a Manhattan building. The suspect was described as 5 feet 6 inches tall and weighing about 160 pounds.
Two months later, police arrested Williams, who stands 6 feet 2 inches tall and weighs 230 pounds. At the time, Williams was driving from his job in Connecticut, where he works with autistic adults, to Brooklyn, with phone data placing him roughly 12 miles away from the incident.
Williams was taken into custody and jailed for more than two days, despite not matching the victim’s description and having little physical similarity to the suspect. The only apparent similarity was that both men were Black with dreadlocks.
“In the blink of an eye, your whole life could change,” Williams tells The New York Times.
A week before the flashing incident, Williams had been photographed for a misdemeanor assault case that was later dismissed. His image remained in NYPD records when investigators examined the February 10 incident. Police reviewed surveillance footage, interviewed the victim, and utilized a facial recognition system, which flagged Williams as a “possible match” but noted that it did not constitute probable cause for arrest.
Investigators then showed the victim six photos of Black men with dreadlocks and facial hair. She selected Williams, who was later proven to have been wrongly identified. Prosecutors dismissed the case last month after Williams’ public defenders from the Legal Aid Society demonstrated his innocence.
“This could have very easily been solved by just really traditional police work,” Diane Akerman, staff attorney with Legal Aid’s Digital Forensics Unit, tells ABC Eyewitness News.
The NYPD has been using facial recognition since 2011, running thousands of searches each year in cases ranging from rapes to murders. In a letter to authorities, Legal Aid highlighted a concerning pattern of false arrests based on facial recognition. They also allege that the NYPD is relying on matches from outside its approved photo database and enlisting other city agencies, like the FDNY, to perform searches it is barred from conducting.
The case is certainly not the first false arrest based on facial recognition technology. Last month, a 51-year-old Florida man found himself the victim of a false arrest after AI facial recognition tech wrongfully accused him of trying to lure a 12-year-old child at a restaurant.
Multiple cities have banned the use of facial recognition technology by police departments, as have some states. However, there is no federal rule against its use. In fact, a report revealed that the FBI had tested widespread facial recognition software on American citizens for almost a decade.
Image credits: Header photo licensed via Depositphotos.
