Cancel the NYPD’s ShotSpotter contract
The NYPD is doubling down on a deadly mistake and starting today, taxpayers will pick up the bill. The department is spending $22 million to expand ShotSpotter, its error-prone, sometimes deadly audio surveillance system — almost doubling the system’s annual cost.
Expanding the system invites tragic shootings by amped up police officers. Police follow ShotSpotter alerts into Black and Brown neighborhoods looking for active shooters who often were never there. Predictably, tragedies ensue. Chicago police chased, shot and killed 13-year-old Adam Toledo after a ShotSpotter alert brought officers to his neighborhood. Rochester police shot Silvon Simmons three times in the back after an alert reportedly first classified as helicopter noise led officers to his backyard. No less than ten NYPD officers fired on and killed an armed man following a ShotSpotter alert in Crown Heights.
We can’t wave off these police shootings as tragic anomalies. Nationwide, ShotSpotter sensors are concentrated in Black and Latinx communities. Police are nearly three times as likely to fatally shoot a man who is Black versus one who is white. Take ShotSpotter in minority neighborhoods, add racist policing, and stir: it’s a deadly cocktail, and the NYPD just ordered a double.
We could charitably interpret NYPD’s move as a response to the COVID-19 crime spike. But shootings were way down before the contract was even finalized. And there’s no proof that ShotSpotter made any difference. The technology didn’t prevent NYC’s uptick in shooting incidents earlier in the pandemic. We certainly can’t credit it with the subsequent drop.
The NYPD’s new ShotSpotter contract primes NYPD officers to do what they’ve done far too often even without this technology: shoot Black and Brown New Yorkers. Yet despite ShotSpotter’s predictable human toll, the evidence suggests that the technology doesn’t do much for public safety. One comprehensive study of ShotSpotter analyzed more than 50,000 ShotSpotter alerts in Chicago. The study shows that ShotSpotter infrequently produces evidence of crimes. It leads to police stops even less frequently. It rarely leads to the recovery of guns. And unsurprisingly, ShotSpotter changes the way police interact with city residents for the worse.
When one looks at how ShotSpotter functions, it’s no surprise. There’s no magic technology that reveals what a sound is. Instead, ShotSpotter uses an array of microphones that run 24/7, recording any noise they hear. When the software thinks there might be a shot (“might” is the crucial word), it’s a human being who decides. Someone sitting in a room hundreds, even thousands of miles away, listening to a random sound from our city, tasked with telling us whether it was a shot…and often getting it wrong.
Even worse, ShotSpotter’s human analysts allegedly change their decisions in response to police pressure. One loud bang in Chicago was reportedly classified by ShotSpotter’s algorithm as a firework, reclassified as a gunshot, and then — months later — “moved” by another analyst to another location to (unsuccessfully) bolster a murder allegation. ShotSpotter’s automated alerts are interpreted by people working allegedly without adequate oversight and under pressure from police.
Our group, S.T.O.P., and other concerned organizations have flagged ShotSpotter’s deal-breaking problems for the NYPD. We told the NYPD that ShotSpotter is driving dangerous police interactions and an increase in illegal stop-and-frisk searches. We highlighted ShotSpotter’s inaccuracy: the tech listens for muzzle blasts, but has been confused by bangs and pops including fireworks, cars backfiring, and even NYC construction noise. We told the NYPD we know that ShotSpotter revises data in response to police requests and ShotSpotter can spy on New Yorkers’ conversations.
The department’s response? Silence. Even after the City Council passed the landmark Public Oversight of Surveillance Technology Act, which requires the NYPD to pull back the curtain and detail how tools like ShotSpotter work, it gave virtually no information on the system’s bias and invasiveness. It denied that ShotSpotter can spy on New Yorkers’ conversations against evidence to the contrary. It refused to supply even the most basic information about ShotSpotter, denying a Freedom of Information Law request for a list of ShotSpotter alerts in NYC, which would allow us to investigate whether ShotSpotter produces any evidence of crimes.
The NYPD has withheld any of the information we would need to approve another three years of ShotSpotter. But what we already know about the technology shows that even $1 would be too much.
If Mayor de Blasio is refusing to reconsider this misguided effort, incoming Mayor Eric Adams and the new City Council must act. It may be a long time before New York is a leader on civil rights, but we can at least follow other cities in dropping ShotSpotter and liberate $7.3 million a year for efforts that actually promote public safety.