Enlarge /. A woman in a mask walks past a Rite Aid store in Portland, Oregon in May 2020.
According to a new report, the pharmacy chain Rite Aid has used facial recognition technologies in hundreds of stores in the country's largest cities, especially in low-income areas where the majority are colored.
Reuters today released a detailed report citing first-hand internal documents, interviews with more than 40 sources familiar with the systems, and camera surveillance in stores. The technology was found to be used in at least 200 stores, including 75 in New York and Los Angeles.
Whenever a customer entered a store that uses the technology, their image was recorded in a database. On repeated visits, the software added new images to existing customer profiles. These images were then compared to a list of "people Rite Aid had previously observed in potential criminal activity." When the software matched, the store's security staff received a push notification for the smartphone.
Rite Aid declined to find out which stores were using the technology, but Reuters journalists in Manhattan and Los Angeles found 33 in 75 stores.
Of these 75 stores, Reuters said, stores in poorer areas were more likely to have facial recognition than shops in higher-income areas – 68 percent of stores Reuters visited in lower-income areas compared 25 percent of business in affluent neighborhoods. In areas where black or Latin American residents were the largest demographic, Rite Aid locations were more than three times more likely to use facial recognition than in predominantly white neighborhoods.
One of the tools that Rite Aid has reportedly used was DeepCam, which is linked to Chinese government mutual funds. Senator Marco Rubio (R-Fla.) Told Reuters the connection was "outrageous," adding, "China's efforts to export its surveillance state to collect data in America are an unacceptable, serious threat."
Rite Aid confirmed the existence of the program to Reuters in February and was defending the use of the technology at that time. A week before Reuters released its story, Rite Aid apparently said the software was no longer used and the cameras themselves were turned off.
"This decision was based in part on a major industry discussion," a Rite Aid representative told Reuters. "Other large technology companies appear to be reducing or rethinking their facial recognition efforts as uncertainty about the benefits of the technology increases."
The "bigger industry talk"
Face recognition technology has notorious differences in effectiveness depending on who it tries to identify. On the whole, the algorithms used work better for men and fair-skinned people than for women and fair-skinned people. Adding masks to the mix that is currently in a pandemic will make facial recognition systems work even worse.
A Detroit-based Rite Aid employee, more than 75 percent of whose population is black, bluntly told Reuters that the software that started the company "didn't take blacks well." The damage prevention employee added: "If your eyes are the same or if you wear your headband like someone else wears a headband, you will get a hit."
The American Civil Liberties Union recently filed a complaint against the Detroit police on behalf of a Michigan man who was arrested in January for a false positive game generated by facial recognition software. In view of the complaint, Detroit chief of police admitted: "If we only used the technology to identify someone, I would say that it is misidentified 96 percent of the time."
The different effects of facial recognition on black people have become part of the discussion amid nationwide protests against police brutality and transgression and in support of black communities. IBM said goodbye to facial recognition business in June. CEO Arvind Krishna said, "AI system vendors and users share a responsibility to ensure that AI is tested for bias, especially when used by law enforcement agencies, and that such bias testing is performed." checked and reported. "Days later, Amazon also imposed a one-year moratorium on police use of the face recognition platform Recognition.