Technology

Gunshot detection firm ShotSpotter expands with new DC office

The acoustic gunshot detection software company ShotSpotter on Wednesday opened its second national incident review center in Washington, D.C., signaling plans for a broader nationwide expansion.

At an event hosted by the firm, Mayor Muriel Bowser (D), D.C. Police Chief Robert Contee and ShotSpotter CEO Ralph Clark discussed how the software can be harnessed by law enforcement to curb gun violence.

“ShotSpotter allows us to make the best use of our police resources,” Bowser said before a ribbon cutting ceremony.

The Metropolitan Police Department (MPD) has deployed ShotSpotter’s technology in the nation’s capital since 2006. The sensors now cover roughly 18 of the city’s 68 square miles, spanning six of D.C’s seven police districts, and the city has paid ShotSpotter roughly $500,000 a year since 2012, according to publicly available purchase records.

Clark told The Hill that the company decided to open an office in D.C. because of its proximity to Capitol Hill as ShotSpotter seeks to expand its operations.

“We’re really proud of the progress we’ve made as a company,” Clark, who took over as the company’s chief executive in 2010, said in a sit-down interview.

ShotSpotter, founded in 1996, works with police departments in more than 120 U.S. cities — as well as a handful of overseas locations — to provide the gunshot detection technology and support staff to analyze sounds.

The new center in D.C. will house a team reviewing gunshot alerts from all over the country, splitting the analysis with the team at the office in Newark, Calif. ShotSpotter is set up to review potential gunshots — through both human analysis and artificial intelligence — within 21 seconds before notifying police departments, according to Ginger Ammon, manager of the Newark operation.

ShotSpotter has been hailed as a way to let police respond to gunfire incidents quickly, especially when there is no 911 call. President Biden last month touted how mayors can purchase gunshot detection systems through the $1.9 trillion American Rescue Plan as a way to “better see and stop gun violence in their communities.”

“Our solutions get officers to the scene of criminal gun fire quickly, precisely and safely,” Clark told the audience at Wednesday’s opening. “They’re more likely to recover physical forensic evidence, interview witnesses and, most importantly, signal to that community that they care. This is a critical component in any gun violence reduction strategy.”

But the technology has received significant criticism from anti-police brutality activists and community organizers in the cities where it is deployed, particularly following the Chicago Police Department killing of 13-year-old Adam Toledo in March.

Officers were deployed to the Little Village neighborhood where Toledo was killed less than five minutes after police received a ShotSpotter notification that eight gun shots had been fired. Rounds were collected from the scene, residents had called 911 about the shots and the police say Toledo was armed when officers arrived.

Clark told The Hill that while the killing was “tragic,” police responded to the alert correctly and the killing has not prompted any changes to ShotSpotter’s business model.

“We want police to respond to the incident — someone was being put in harm’s way,” Clark said.

Concerns about ShotSpotter from critics have focused on the accuracy of the technology.

ShotSpotter says it has a 97 percent accuracy rate, determined via feedback from its customers, that takes into account detection rate, location accuracy and speed.

That figure has been challenged however, with critics saying the technology is prone to false positives from things like fireworks or cars starting.

Jonathan Manes, an attorney and adjunct professor at Northwestern Law, released a report earlier this year analyzing ShotSpotter alerts and police reports from Chicago over an 18-month period that found 89 percent of deployments initiated by the detection technology turned up no gun-related crime.

“It’s hard to see it being worth … chasing down ShotSpotter alerts when such a small proportion actually produces results,” Manes told The Hill.

Clark pushed back on the study, arguing that the lack of a police report does not necessarily mean there isn’t a gunshot.

Independent analyses of the accuracy are limited.

ShotSpotter makes efforts to weed out false alerts using machine and human oversight.

Ammon, the Newark incident center manager, told The Hill that reviewers are trained for several weeks on a variety of sounds to learn how to distinguish gunfire. The company employs 21 reviewers who are assigned shifts based on historical crime frequency.

ShotSpotter sent out roughly 240,000 gunshot alerts in 2020, Clark said.

Verifying accuracy data for D.C. is difficult. The city releases a public list of ShotSpotter alerts quarterly, but those releases do not include what ultimately happened after police were deployed.

Another major concern for critics of ShotSpotter is that it can justify over policing of Black and brown communities.

“Technology that is given to MPD is basically almost always used to harass, follow and assault people of color in D.C.,” Valerie Wexler, a spokesperson for the Stop Police Terror Project D.C., told The Hill.

Wexler pointed to MPD’s gang database as an example of how technology can be used to justify the need for policing certain communities in the first place.

“The gang database created this kind of endless circle by naming anyone who hangs out with someone they call a gang member also a gang member,” she said. “It’s the same thing with tools like ShotSpotter: ShotSpotter heard gunfire so there must be a lot of crime, because there’s a lot of crime we need to pour even more money into ShotSpotter and surveillance.”

Clark said that coverage decisions are made by local police departments based on “objective data” and gunshot and crime rates.

The only police district in D.C. not covered by ShotSpotter sensors is composed primarily of two city wards that have a high concentration of white residents.

There are also privacy concerns with the technology. ShotSpotter carefully limits the amount of audio that it collects and the time that it holds that audio.

The company had an independent privacy audit conducted by the nonprofit Policing Project at the New York University School of Law in 2019 which concluded that “the risk of voice surveillance is extremely low.”

Still, experts worry that in conjunction with other police surveillance technology in D.C., including facial recognition, cell site simulators and license plate readers, ShotSpotter can reduce privacy.

“These technologies really certainly pose clear threats to our privacy, especially when used in an interconnected way, they can sort of map our whereabouts,” Lauren Sarkesian, senior policy counsel at New America’s Open Technology Institute, told The Hill.

Open Technology Institute is part of a coalition of local groups headed by the American Civil Liberties Union of D.C. called Community Oversight of Surveillance seeking to end unchecked surveillance of local residents and involve the community in making decisions over how technology is acquired and used in the area.

“A lot of these technologies are very secretly acquired and deployed on citizens like you and me, and people just don’t have awareness of not only that they’re abused but what the impact is,” Sarkesian said.