ShotSpotter CEO says technology can reduce gun violence
Toronto mayor considers surveillance device that has mixed reviews
Technology the Toronto Police Board wants to use to reduce gun violence may not be effective, says a former U.S. police officer.
ShotSpotter is designed to pinpoint the location of shootings as they happen.
Ralph Clark, CEO of ShotSpotter, heralds the technology for its accuracy and immediacy. It can inform police within 30 to 45 seconds after a trigger is pulled where the gun shot is coming from.
Sensors are designed to suppress ambient noise so the device can detect “pops, booms, or bangs” emanating from the muzzle of a gun, Clark said.
When the sensors recognize the noise, it’s timestamped and analyzed to triangulate where the shot originated. Sensors are typically located on rooftops
More than 90 cities in the U.S. are using the technology, Clark said. The devices are located in areas with high instances of gun crime. Mayor John Tory has proposed buying the technology, along with 40 additional sur- veillance cameras. Tory said he wants to put in the request for funding next week, when city council holds its last meeting before December.
New York City implemented ShotSpotter in 2015 to stem gun violence. In written statement to the Star, detective Sophia Mason, of the NYPD, called it a “highly effective crime fighting tool that helps law enforcement respond to shootings quicker and investigate them more precisely than ever before.
“Crime is at historic lows in our city thanks to the NYPD’s innovative precision and neighbourhood-policing strategies, which utilize cutting-edge technology like ShotSpotter.”
Not everyone is convinced of the utility of the technology. Larry Smith, a former Baltimore police officer of 18 years, said there police should not rely on it to thwart gun violence; the solution should come from the community, in the form of providing more job opportunities for disenfranchised youth. The technology is used in Baltimore.
Smith said he has concerns over the privacy of the information collected. “The company owns the data, so I don’t know what they do with it, and they’re collecting the alerts and all the statistics, but they don’t release and share it,” he said.
Smith said the police in Baltimore are “very” short staffed, and he’s concerned officers are being led astray, responding to false alarms as detected by ShotSpotter.
“You’re already spread thin and you’re already responding to every single ShotSpotter alert. I just don’t know how practical that’s going to be,” said Smith, who added that he hasn’t learned of any arrests being made that can be attributed to the technology. ShotSpotter was founded about 20 years ago.
“We have acoustic experts that not only listen to these recording snippets, but they’re also looking at how the sensors participated, they’re looking at the waveform file ... and they’re making the final determination if that’s a gunshot or not,” said Clark. The information is then routed to subscribing police agencies, he continued, in the form of a digital alert, which shows a precise location on a map and other details, such as the number of rounds fired.
“At least three sensors have to detect an the pop, boom, or bang for us to create an incident,” said Clark.
Clark said 90 per cent of outdoor gunfire is detected. He added that misclassification is negligible. Location accuracy is within 25 metres, he said.
“People aren’t concerned about the false positives so much; they’re concerned about the false negatives, like, ‘wait a minute, I have confirmed gunshot here and you didn’t detect it.’ We don’t want to not publish something that was a gunshot,” he said.
In the U.S. the technology costs between $65,000 to $90,000 USD per 2.6 square kilometres covered.
Clark downplayed privacy concerns, saying the technology cannot pickup on voices and people’s whereabouts.
The device is always on, listening, he said, but it has limited memory.
“When they hear the pop, boom, or bang, that’s the thing that gets sent to the brains and the memory. Everything else is at the sensor level and it gets written over. The only thing the agencies ever see is gunshots,” said Clark, who noted that anything else is filtered out by both machine and human operators.
“This is very narrow type of surveillance.”