Facial Recognition Software Leads to Mistaken Arrest of Georgia Man

Anybody who has seen a Hollywood dystopian movie knows that the more automated our lives become, the closer we get to living in underground cities, while a few elite humans fight the machines for dominance.

Artificial intelligence has a sinister side effect when it comes to law enforcement. After Americans discovered that San Francisco has a killer robot program, San Francisco suspended its killer robotics program. Robotic aids are similar to those used to contain and examine bombs. However, they were approved for use against violent suspects in “extreme circumstances.” It seems that sending a robot into a human situation to evaluate and respond to it is a good way to make very poor mistakes. Robotic aids cannot be trusted and can’t replicate human judgment and interaction.

A Georgia man was recently arrested in Louisiana on a warrant for facial recognition technology. Randall Reid (28), was arrested in DeKalb County Georgia last November. Authorities had linked him to a series of purse thefts in Jefferson Parish, Louisiana, and Baton Rouge.

Randall claimed he had never been to Louisiana and that he didn’t know what “Jefferson Parish” was. It was impossible for him to have done it. Problem is, he said he did.

“They said that I had a warrant from Jefferson Parish. Reid replied, “What is Jefferson Parish?” “I’ve never been to Louisiana in my entire life. They then told me that it was theft. They also said that I have not been to Louisiana and I don’t steal.

Baton Rouge authorities issued an arrest warrant after facial recognition software linked surveillance images with Reid’s Georgia identification records. The warrant was executed by Georgia authorities and the man was taken into custody.

After noticing significant discrepancies between them, Reid was eventually released. Reid had a mole on his face while the suspect didn’t. The men also had a difference of at least forty pounds. They shared one thing in common: they were both black.

This fact has reaffirmed concerns about the inherent danger of relying upon technology in law enforcement situations. Facial recognition software has been known to misidentify blacks and other minorities at a higher rate than whites.

A MIT study found that three commercial gender recognition systems had error rates of up to 34% in dark-skinned females. This is nearly 49 times the rate for white males.

Similar findings were revealed by a Commerce Department study that was completed late last year. The study examined instances where an algorithm incorrectly identified two people as the same person. It found that the error rates for African men or women were two orders higher than those for Eastern Europeans.

The technology isn’t only creepy because of racial vulnerabilities. It’s also very creepy. A New York City lawyer was recently turned away from the holiday Rockettes performance because facial recognition technology in the lobby identified that she was a lawyer representing the parent entertainment company Madison Square Garden Entertainment. After being flagged by software, the woman was only going to be with her daughter’s Girl Scout troop for an afternoon of fun.

Although one might argue that the show was entitled to reject anyone who could be gathering information for a lawsuit, using facial recognition software to do this creates more problems than it solves. This is especially true when it comes to privacy.

China already uses facial recognition in China to identify citizens in almost every area of their lives. This alone should give Americans cause for concern when it comes to technology.