by Douglas Ankney
According to a report by sciencefriday.com, one in every two American adults is in a law enforcement facial recognition network. Most adults have unwittingly consented to the release of their photos that they have uploaded to social media, including dating sites.
While it’s impossible to determine the exact number of people in a facial recognition database, the Georgetown Law Center on Privacy and Technology reported that over 117 million people are in law enforcement facial recognition networks.
In addition, the Government Accountability Office found that in a four-year period the FBI conducted over 118,000 face recognition searches on its database.
U.S. Immigration and Customs Enforcement has mined the Department of Motor Vehicles databases of states that grant driver’s licenses to undocumented immigrants. And numerous police departments scan the faces of passersby using hand-held surveillance cameras.
But the accuracy of facial recognition technology is questionable. While Chinese authorities managed to locate and arrest a man in a crowd of 60,000 at a concert using recognition technology, Amazon’s software “Rekognition” misidentified 28 members of Congress and matched them with criminal mugshots. The rate of ID mismatch is far worse for women and people of color. Facial recognition can be thought of as a “faceprint.”
A computer analyzes each person’s unique facial structure, such as the distance between both eyes and the distance from the nose to the lips. It then maps those key features onto an existing image, or compares them against a database of existing images.
Apparently, these measurements are subject to change with varying facial expressions, which might account for some of the mismatches. Additional misidentification factors include jewelry, clothing, hairstyles and lighting.
As a digital subscriber to Criminal Legal News, you can access full text and downloads for this and other premium content.
Already a subscriber? Login