Some recent test results out of New York City have highlighted some of the limitations of facial recognition technology. The Metropolitan Transit Authority had been hoping to use facial recognition cameras to identify potential terrorists driving through the city, and set up a pilot program to test the system at bridge and tunnel crossings.
Unfortunately, the MTA’s system was unable to recognize a single face “within acceptable parameters” during testing on the Robert F. Kennedy Bridge, according to a leaked internal email, indicating that the intrusive technology hasn’t been of much help to the city’s anti-terrorism efforts.
The results suggest that some biometric surveillance technologies simply aren’t ready for certain real-world settings. While previous testing has been 80 percent accurate when identifying faces through a windshield, those tests were conducted in low-speed environments. Drivers going about their day generally move at much higher speeds, even when dealing with New York City traffic.
The findings also raise concerns about the reliability of facial recognition tech, and could further erode the public trust after the exposure of the NYPD’s secret surveillance program. At the very least, this will likely give fuel to activists worried about the loss of privacy and other potential abuses of the technology, especially in cases of government and police surveillance.
The MTA has indicated that its pilot program will continue, but it will have to post much more accurate results if it wants to win over skeptical citizens.
April 9, 2019 – by Eric Weiss