xml:space="preserve">
xml:space="preserve">
Advertisement
Advertisement

Another arrest, and jail time, because of a bad facial recognition match

The Hampton Inn in Woodbridge, N.J. on Dec. 28, 2020. Nijeer Parks was accused of shoplifting and trying to hit an officer with a car and is the third known Black man to be wrongfully arrested based on face recognition.
The Hampton Inn in Woodbridge, N.J. on Dec. 28, 2020. Nijeer Parks was accused of shoplifting and trying to hit an officer with a car and is the third known Black man to be wrongfully arrested based on face recognition. (Mohamed Sadek/The New York Times)

In February 2019, Nijeer Parks was accused of shoplifting candy and trying to hit a police officer with a car at a Hampton Inn in Woodbridge, New Jersey. He had been identified by police using facial recognition software, even though he was 30 miles away at the time of the incident.

Parks spent 10 days in jail and paid around $5,000 to defend himself. In November 2019, the case was dismissed for lack of evidence.

Advertisement

Parks, 33, is now suing the police, the prosecutor and the city of Woodbridge for false arrest, false imprisonment and violation of his civil rights.

He is the third person known to be falsely arrested based on a bad facial recognition match. In all three cases, the people mistakenly identified by the technology have been Black men.

Advertisement
Advertisement

Facial recognition technology is known to have flaws. In 2019, a national study of over 100 facial recognition algorithms found that they did not work as well on Black and Asian faces. Two other Black men — Robert Williams and Michael Oliver, both of whom live in the Detroit area — were also arrested for crimes they did not commit based on bad facial recognition matches. Like Parks, Oliver filed a lawsuit against the city over the wrongful arrest.

Nathan Freed Wessler, an attorney with the American Civil Liberties Union who believes that police should stop using face recognition technology, said the three cases demonstrate “how this technology disproportionately harms the Black community.”

“Multiple people have now come forward about being wrongfully arrested because of this flawed and privacy-invading surveillance technology,” Wessler said.

He worries that there have been other arrests and even mistaken convictions that have not been uncovered.

Advertisement

Law enforcement often defends the use of facial recognition, despite its flaws, by saying that it is used only as a clue in a case and will not lead directly to an arrest. But Parks’ experience is another example of an arrest based almost solely on a suggested match by the technology.

The crime

On a Saturday in January 2019, two police officers showed up at a Hampton Inn in Woodbridge, after receiving a report about a man stealing snacks from the gift shop.

The alleged shoplifter — a Black man, nearly 6 feet tall, wearing a black jacket — was visiting a Hertz office in the hotel lobby, trying to get the rental agreement for a gray Dodge Challenger extended. The officers confronted him and he apologized, according to the police report. He said he would pay for the snacks and gave the officers a Tennessee driver’s license.

When the officers checked the license, they discovered it was fraudulent. According to a police report, one of the officers spotted a “big bag of suspected marijuana” in the man’s pocket. They tried to handcuff him. That’s when the man ran, losing a shoe on the way to his rental car, police said.

As he drove off, the man hit a parked police car and a column in front of the hotel, police said. One of the officers said he had to jump out of the way to avoid getting hit. The rental car was later found abandoned in a parking lot a mile away.

The match

A detective in the Woodbridge Police Department sent the photo from the fake driver’s license to state agencies that had access to face recognition technology, according to a police report.

The next day, state investigators said they had a facial recognition match: Parks, who lived in Paterson, 30 miles away, and worked at a grocery store. The detective compared Parks’ New Jersey state ID to the fake Tennessee driver’s license and agreed it was the same person. After a Hertz employee confirmed that the Tennessee driver’s license photo was of the shoplifter, police issued a warrant for Parks’ arrest.

“I don’t think he looks like me,” Parks said. “The only thing we have in common is the beard.”

Parks’ mistaken arrest was first reported by NJ Advance Media, which said that the facial recognition app Clearview AI was used in the case, based on a claim in Parks’ lawsuit. Parks’ lawyer, Daniel Sexton, said he had inferred that Clearview AI was used, given media reports about facial recognition in New Jersey, but now believes he was mistaken.

Clearview AI is a facial recognition tool that uses billions of photos scraped from the public web, including Facebook, LinkedIn and Instagram. Clearview AI’s founder, Hoan Ton-That, said officers affiliated with the state agencies where information was analyzed in the case, known as fusion centers, involved in the case were not using his company’s app at that time.

According to the police report, the match in this case was to a license photo, which would reside in a government database, to which Clearview AI does not currently have access. The law enforcement involved in making the match — the New York State Intelligence Center, New Jersey’s Regional Operations Intelligence Center and two state investigators — did not respond to inquiries about which facial recognition system was used.

In January, after a New York Times story about Clearview AI, New Jersey’s attorney general, Gurbir S. Grewal, put a moratorium on Clearview’s use by police and announced an investigation into “this product or products like it.” A spokesman for the attorney general’s office said New Jersey’s Division of Criminal Justice is still evaluating the use of facial recognition products in the state, and that the development of a policy governing their use is ongoing.

‘I was afraid.’

After police arrested Parks, he was held for 10 days at the Middlesex County Corrections Center. New Jersey’s no-bail system uses an algorithm that evaluates the defendant’s risk rather than money to determine whether a defendant can be released before trial.

A decade ago, Parks was arrested twice and incarcerated for selling drugs. He was released in 2016. The public safety assessment score he received, which would have taken his past convictions into account, was high enough that he was not released after his first hearing. His mother and fiancée hired a private attorney, who was able to get him out of jail and into a pretrial monitoring program.

His prior history with the criminal justice system is what made this incident so scary, he said, because this would have been his third felony, meaning he was at risk of a long sentence. When the prosecutor offered a plea deal, he almost took it even though he was innocent.

“I sat down with my family and discussed it,” Parks said. “I was afraid to go to trial. I knew I would get 10 years if I lost.”

Parks was able to get proof from Western Union that he had been sending money at a pharmacy in Haledon, New Jersey, more than 30 miles away, when the incident happened. At his last court hearing, he told the judge he was willing to go to trial to defend himself. But a few months later, his case was dismissed.

Robert Hubner, the chief of the Woodbridge Police Department, declined to comment on the case because of the pending lawsuit but said his department had not been served the complaint. The Middlesex County prosecutor’s office also declined to comment.

Parks’ lawsuit over the wrongful arrest does not yet ask for damages.

Advertisement

“I was locked up for no reason,” Parks said. “I’ve seen it happen to other people. I’ve seen it on the news. I just never thought it would happen to me. It was a very scary ordeal.”

Advertisement

Jack Begg contributed research.

c.2020 The New York Times Company

Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement