“Glorified Roomba that narcs.” “Minority Report, but instead of three pale orphans hanging out in goo, it’s a bunch of bland-looking computers.” The Last Week Tonight With John Oliver team came up with a lot of zingers to describe police surveillance tools, but none is more apt or chilling than, as one official puts it, “Big Brother on steroids.”
The show’s deep dive in the August 2 episode scanned the country’s “vast ecosystem” of police surveillance tech, which it notes has become as pervasive as it is invasive. So Oliver and his writers put together a “field guide” to the five most prominent methods or tools: cameras, microphones, automated license plate readers, cell site simulators, and predictive policing (the last of which earned some high praise in its infancy from Time magazine). The first one is probably the most common/commonly sighted, but the actual stats are staggering: In addition to all of the cameras owned and operated by various law enforcement agencies, there are an estimated 85 million privately owned cameras in U.S., which can be accessed with just a click thanks to businesses like Fusus.
But thanks to systems like Shotspotter, microphones are also popping up all over cities, ostensibly to alert police to sounds that could be gunfire but have so far proven less than effective (though, you might argue that’s exactly the point). Automated license plate readers are also spreading, with businesses like Flock setting up these ALPRs throughout cities big and small and feeding the data to police departments, no warrant required. And speaking of jaw-dropping stats, Oliver cites a study by the Brennan Center For Justice about the use of ALPRs in Vallejo, California, that found that 37% of all ALPRs “hits” from fixed readers (such as those attached to street lights) and 35% of mobile ALPRs were misreads, which is “an astonishingly high error rate.”
Cell site simulators like Stingray, which allows cops to snoop on any nearby phones, are also catching on. These “man-in-the-middle attacks” trick cell phones into thinking they’re connecting to the nearest cell tower, giving police access to personal info. Like the other surveillance tools in the segment, they don’t work in isolation—they feed that data to real-time crime centers.
Last but certainly not least invasive is predictive policing, which Oliver calls the “dumbest” of all of these methods. Predictive policing was developed in Santa Cruz, CA, with the aid of a researcher at the University Of Santa Clara, who used a lot of earthquake terminology to describe the approach. As Oliver notes, these systems or devices “are disproportionately found in communities of color,” and when used alongside things like facial recognition software, which has already been shown to misidentify people based on race, have a clear racial bias. Santa Cruz ultimately backed away from predictive policing for that reason, but as Oliver notes at the start of the segment: “Technology is amazing, but it does feel like turning on the panopticon should be a bit harder than that.”
Keep scrolling for more great stories.