Man Sues Sunglass Hut After Faulty AI Facial Recognition Tech Led to Wrongful Arrest


Sunglass Hut | Belinda Jiao/ZUMAPRESS/Newscom

Harvey Murphy was arrested in October 2022 for the armed burglary of a Huston-area Sunglass Hut store after a facial recognition device identified him as the burglar. The only problem? Murphy was in Sacramento, California at the time of the crime—thousands of miles away.

Making matters worse, Murphy, now 61, says he was brutally sexually assaulted in jail just hours before he was set to be released after the charges against him were dropped. In a lawsuit filed last week, Murphy claims that his arrest was the result of gross negligence from the facial recognition company—and he demands $10 million in damages to compensate for his wrongful imprisonment.  

On January 22, 2022, two men entered a Sunglass Hut store in Huston, Texas, with one threatening the store employees with a gun and forcing them into a storeroom at the back of the business. The two intruders eventually left with the money in the store’s cash register as well as several pairs of sunglasses. 

According to the complaint, local police began investigating the crime as soon as they arrived on the scene. Soon after, Anthony Pfleger, the head of loss prevention at EssilorLuxottica, Sunglass Hut’s parent company, called the police and told them that the armed burglar had been identified. Apparently, facial recognition software had concluded that Harvey Murphy was the criminal—and that he had even been responsible for additional robberies of a different Huston-area Sunglass Hut and a Macy’s store.

Police then presented a photo lineup to one of the store employees, who identified Murphy as the robber. However, the complaint alleges that the clerk had been “prepped” by EssilorLuxottica, leading her to incorrectly identify Murphy. Unbeknownst to Murphy, a warrant was issued for his arrest.

In October 2022, Murphy, who had been living in California, returned to Texas for work. After attempting to renew his driver’s license at a local DMV, he was arrested and booked into the Harris County jail. According to the complaint, once Murphy learned he was being charged for the January 22, 2022 robbery, he informed his lawyer that he was in a different state at the time of the crime. Soon after, the District Attorney agreed to drop the charges against Murphy.

However, just hours before Murphy was set to be released, he says he was beaten and gang raped in a jail bathroom by three prisoners. According to Murphy’s lawsuit, after being brutally attacked, his assailants threatened to kill him if he reported the crime. 

“Murphy was released a few hours later, but in some ways, he never left jail,” the complaint writes. “His time in jail will stay with him forever. Not an hour goes by without Murphy relieving the brutal attack and rape. Worse than that, the attack left him with permanent injuries that he has to live with every day of his life. All of this happened to Murphy because the Defendants relied on facial recognition technology that is known to be error prone and faulty.”

This is far from the first time that facial recognition technology has led to wrongful arrests. In February 2022, a Detroit woman who was eight months pregnant was falsely arrested for robbery and carjacking. That same year, an Atlanta man was falsely arrested for a series of thefts in Lousiana, a state the man had never been to. In November 2023, Miami police used facial recognition technology to arrest a homeless man who had refused to give his name to an officer.

“What companies like EssilorLuxottica and Macy’s tell the public is that facial-recognition systems can have near-perfect accuracy,” Murphy’s suit states. “What they do not tell the public is that ‘near-perfect accuracy’ only occurs in absolutely ideal conditions … in the real world, accuracy rates are much lower.”

The post Man Sues Sunglass Hut After Faulty AI Facial Recognition Tech Led to Wrongful Arrest appeared first on Reason.com.

from Latest https://ift.tt/Ch2q0rJ
via IFTTT

Leave a Reply

Your email address will not be published. Required fields are marked *