Can Facial Recognition Software Lead to False Arrests?

Experienced federal criminal defense lawyer John Helms discusses the rise in facial recognition software and how it can lead to a false arrest.

If you haven’t heard of facial recognition software, it’s probably only a matter of time. Widely used in China, facial recognition is a type of technology that uses cameras and artificial intelligence to track people and identify faces.

Although the technology potentially has a wide range of uses, in China it’s used primarily by police, which relies on it to surveil citizens and help identify perpetrators of crime. In many cases, police in China even use the technology to detect crimes in progress and pinpoint the individuals responsible for the offenses.

The problem, however, is that despite having “intelligence” in its name, the artificial intelligence behind facial recognition software isn’t foolproof. In some cases, it has misidentified people, leading to false arrests. This problem has prompted many criminal justice experts to issue warnings about possible civil rights violations and false prosecutions.

The Rise of Facial Recognition in China

According to a New York Times report, China has about 200 million surveillance cameras spread throughout the country, which is four times the number in the United States.

The cameras employ facial recognition software that is used in a multitude of different ways. In some cases, cameras broadcast the faces of people caught jaywalking. In other cases, cameras show the faces and identities of people accused of not paying their debts. Different cameras require people to submit to facial scanning before they can gain access to a housing unit.

In some parts of China, police officers even wear special smart glasses equipped with artificial intelligence software that can scan faces. In popular tourist spots, tall poles mounted with dozens of cameras loom above the public areas. The government also says it would like to increase the number of surveillance cameras to 300 million by 2020.

Amazon’s Facial Recognition Software Comes Under Fire for Mix-ups

China isn’t the only country currently experimenting with facial recognition software. In the United States, tech giant Amazon has developed similar surveillance technology. In July 2018, however, the company drew ire when its software, which goes by the name “Rekognition,” incorrectly identified 28 members of Congress as individuals arrested for crimes.   

According to reports, the Rekognition software disproportionately falsely targeted people of color in the mix-up. In the test, which was conducted by the American Civil Liberties Union (ACLU), researchers ran Rekognition searches on all 535 members of Congress and compared them against 25,000 mugshots.

Other reports note that similar facial recognition software also suffers from a poor track record. The software used by police in metropolitan London returns 49 false matches for every result, which means police must manually review every false match.

In one report, facial recognition technology used by police at a soccer match in Wales resulted in a 92 percent false positive rate. Among the 2,470 alerts the software returned, an astounding 2,297 were bad matches. This statistic means that nine out of 10 matches were incorrect.  

Bad Facial Recognition Technology That Leads to False Arrests

In the U.S., facial recognition technology has already led to false arrests, excessive use of force, and malicious prosecution. In 2014, a Denver man was arrested for two bank robberies after police relied on facial recognition software used by the FBI.

The problem was the man, who worked as a financial advisor at the time, didn’t commit either robbery. The facial recognition software failed to spot a mole on the innocent man’s right cheek.

The software also matched the man with another individual even though the innocent man was three inches taller than the person who actually committed the robberies.

When police arrested the Denver man, he says they used flash-bang grenades that caused him to develop a post-traumatic stress disorder. He also showed reporters where police allegedly knocked out one of his teeth.

Additionally, the man says the false arrest and prosecution resulted in him not seeing his children over a period of two years. He also lost his job due to the misidentification and subsequent arrest. He also suffered cracked ribs and other injuries when police tackled him to the ground outside his home.

Ultimately, the man’s defense lawyers produced records from his employer showing that he was at work and on the phone at the time of the robbery. Later, police discovered that the two robberies had actually been carried out by two different men — neither of whom was the innocent man.  

According to a report, the man sued police for $10 million over the excessive force, medical trauma, and other damages that stemmed from it.

The case is obviously an extreme example of police misconduct, but it highlights the very real dangers of facial recognition technology. If you have been wrongly accused of a crime, you should speak with an experienced Texas criminal defense lawyer as soon as possible.

Media Contact:

Dallas Criminal Defense Attorney John Helms

(T): 214-666-8010