A recent New York Times article details the frighteningly pervasive application of facial recognition softwares, originally developed for overseas military use, to local police departments across the country. Because the application of the technology to domestic police forces is rather recent, there are no clear guidelines or restrictions on its use. Timothy Williams reports:
“The software can identify 16,000 points on a person’s face – to determine the distance between the eyes or the shape of the lips, for instance – and compare them with thousands of similar points in police booking or other photos at a rate of more than one million faces a second.”
Despite public reservations about police misconduct, “the F.B.I. is pushing ahead with its $1 billion Next Generation Identification program, in which the agency will gather data like fingerprints, iris scans and photographs, as well as information collected through facial recognition software. That software is capable of analyzing driver’s license photos and images from the tens of thousands of surveillance cameras around the country. The F.B.I. system will eventually be made accessible to more than 18,000 local, state, federal and international law enforcement agencies,” the article reports.
“But people who are not criminal suspects are included in the database, and the error rate for the software is as high as 20 – meaning the authorities could misidentify millions of people.”
Such cutting edge technologies that merge biometric measurements with the power of big data draw on a long history. As early as 1879, French police officer and researcher Alphonse Bertillon devised a series of standardized measurements through which criminal suspects could be identified. Introduced in the US in 1887, Bertillon System would gain widespread acceptance until it was supplanted by fingerprinting as a primary means of law enforcement identification.
Bertillon’s reliance on facial measurements was contemporaneous with the popularization of phrenology and later eugenics, which used similar measurement techniques to attempt to distinguish superior and inferior biological “races”. How might the racial application of such measurements have distorted its use by law enforcement? How might contemporary issues of racism in policing be impacted by the use of new biometric technologies?
The expanding use of biometric technologies (facial recognition, fingerprinting, DNA recognition, palm vein pattern recognition…) by government and law enforcement agencies poses new questions at the intersections of ethics, technology, policing, and surveillance. Far from some imagined future police state, the real-world applications are taking place here and now. Dubai police are equipped with Google Glass with facial recognition software to help track down suspects. Meanwhile, the F.B.I. is rolling out a $1.2 billion “Next Generation Identification” system, equipping 62 police departments with handheld iris and facial recognition devices.
Enter artist Zach Blas, who has designed a series of aesthetically captivating masks that serve the practical purpose of shielding subjects from potential facial recognition technology, while making a poignant political statement against unethical government policing and surveillance.
“I think one of the things that you’re seeing, when you ask about the potential ramifications of this, is you’re seeing a really odd return to pseudo-scientific endeavors of the 19th century.
Things like anthropometry, physiognomy, phrenology, methods that were popularized by Alphonse Bertillon, Cesare Lombroso, Francis Galton, certain criminologists who would measure criminal skulls to say, “Here’s what the average criminal skull looks like.”
You’re seeing a really weird history-repeating-itself moment with things like that coming back, but of course they’re executed under the guise of high-tech biometrics, so they’re not as questioned because there’s this high-tech sheen to these technologies that I think a lot of people believe are utterly scientifically objective. I think that’s one of the central problems, that biometrics propagate a certain way of understanding identity where you can scan the surface of the body digitally, fully quantify it, and gain some kind of core truth about a person. And you now see that permeating in a lot of different ways.”