A 61-year-old man has sued department store Macy’s and Sunglass Hut for using a facial recognition system against him to falsely accuse him of robbing their stores at gunpoint. Harvey Eugene Murphy Jr., as the complainant is called, was arrested in October 2023 for the theft from one of these stores in the Houston area (United States). While in custody and until the alibi proving his innocence was proven, Murphy was raped and beaten, according to his complaint. If this is confirmed, it would be a new case in which facial recognition systems applied to citizen security cause errors with serious consequences.
The fateful story begins in January 2022, when two armed robbers steal thousands of dollars and products from one of these Houston eyewear stores. Nearly two years later, police arrested Murphy while he was in Texas at an office renewing his driver’s license. According to the complaint that was filed Guardianthe arrested person claims that it was an employee of Sunglass Hut’s parent company (EssilorLuxottica), in collaboration with Macy’s, who gave his name to authorities.
As Houston police investigated the armed robbery, that employee used facial recognition software to identify Murphy as one of the robbers. Even though the image entered into the system came from poor-quality cameras, according to the lawsuit, the EssilorLuxottica representative called the police to inform them that he had already identified one of the culprits using this technology. Additionally, he accused Murphy, who was then living in California, of being responsible for two other thefts.
Murphy is now demanding $10 million (a little over nine million euros) from Macy’s and EssilorLuxottica for the misuse of this facial recognition system which wrongly identified him. During his ten-day detention in Harris County Prison (Texas), he suffered beatings and gang rape, which left him with serious physical and psychological consequences.
“The attack left him with permanent injuries that he must live with every day of his life,” the lawsuit states. “All of this happened to Murphy because the defendants relied on facial recognition technology, which is known to be error-prone and faulty.”
Murphy’s case once again highlights the shortcomings and dangers of relying on facial recognition, particularly in an area as sensitive as the prosecution of suspected criminals, as it is known to produce false positives, particularly against Black. His lawyer, Daniel Dutko, explained to local media that this technology is the only plausible explanation for his false identification and arrest.
“We know that studies show that African Americans have a high rate of false positives based on facial recognition software. We also know that comparing older people with photos of themselves when they were younger results in a very high false positive rate of up to 90%,” says Dutko. Murphy had a criminal record dating back to the 1980s and 1990s, which is why he showed up in the system that misidentified him. This would be the first known case of a white man being wrongly accused of using the technology. “Mr. Murphy’s story is disturbing to every citizen of this country,” Dutko said in a statement. “Anyone could be wrongly accused of a crime based on error-prone facial recognition software, just like him.”
In recent years, several cases have warned of the danger of trusting these algorithms to arrest citizens, as was the case with Robert Julian-Borchak Williams, arrested by the Detroit police, or that of Porcha Woodruff, who was pregnant and the real thief was not. It is for this reason that the use of facial recognition to fight crime has been one of the most controversial points in the debate over the pioneering law on artificial intelligence in the European Union.