What is ShotSpotter? Controversial gunshot detection technology facing increasing scrutiny

Last week, both Massachusetts senators called on the federal government to investigate its funding of a controversial gunshot detection system in the state, saying it may violate civil rights. 

In March, Somerville’s councilor at large introduced an order for the city to end its use of the same technology after employing it for a decade. 

A multitude of cities across the nation — including Chicago — recently decided not to renew their contracts for the system.

The tool in question is ShotSpotter — an audio recording technology designed to detect gunshots and send officers to a crime scene in minutes. 

The relatively low-profile system has divided police and civil liberty activists in recent years, part of ongoing debates about police surveillance, potentially discriminatory tactics, and the use of force and technology in policing. 

Critics say the system contributes to over-policing of minority communities and is largely inaccurate, sending officers on high alert to areas with already-fraught relationships with police.

Police say the system is life-saving, helping first responders get to crime scenes quickly in situations when people may not call 911. 

The debate over whether to use, or keep using the system has taken center stage in town and city council meetings across the country. But the relative lack of data about the technology and complicated studies that show contrasting results has made for a difficult decision for municipalities. 

“Taking a beat to try to figure out ‘What is our internal data, are we looking at it? And how can it better inform public policy?’ … seems to me an appropriate way to handle a situation like this where there have been such tight constraints on city data,” said Katy Naples-Mitchell, director of the Program in Criminal Justice Policy and Management at Harvard’s Malcolm Wiener Center for Social Policy. “It’s always a good idea, especially when there are such significant civil rights and community concerns on the table.” 

What is ShotSpotter? How does it work?

The company ShotSpotter, now known as SmartThinking, was founded in 1996 by a physicist and two engineers, using technology inspired by physicist Robert Showen’s work to determine the location of earthquake epicenters. The company’s product is now used in more than 160 cities, according to its website.

The technology itself employs acoustic sensors that are placed high above the street to avoid street-level sounds, according to Tom Chittum, senior vice president of Forensic Services at SoundThinking. The sensors are triggered by loud sounds that resemble gunshots. 

An alert is triggered if three separate sensors detect a loud sound. The system then calculates the time the sound reached individual sensors to determine the location of the gunshot, Chittum said. 

An algorithm determines whether the sound is consistent with a gunshot. The recording then goes through further review from employees at SoundThinking’s incident review center. The company says the entire process takes less than a minute between the shots fired incident to showing up in a department’s call center or on an officer’s computer or phone. 

In order to determine where sensors go, police departments “analyze and provide historical gunfire and homicide data” to the company. The microphones are then placed in areas “in most need of gunshot detection,” the company said in a statement to Boston.com. 

But there’s an element of secrecy to the system — neither police departments nor cities and towns know where each individual sensor is located. SoundThinking says this confidentiality is to protect community members.

The system has been used in the state’s largest city, Boston, since 2007. The city paid about $1.5 million to install the system, WBUR reported in 2011. 

Cambridge, Worcester, Chelsea, Somerville, and Lawrence are among the Boston-area cities that have contracts with SoundThinking, according to their respective police departments. 

ShotSpotter says its accuracy is 97%, contradicting other studies 

The system’s accuracy is one of its most controversial aspects, with different studies coming to wildly varied conclusions. 

SoundThinking insists that it has a 97% accuracy rate — with a 0.5% false positive rate — based on an independent audit of the system by Edgeworth Economics. The company commissioned the study after the MacArthur Justice Center (MJC) published their own study in 2021 using data from Chicago that indicated 89% of the system’s triggers turned up no gun-related crime. 

In its response to the MJC study, Edgeworth wrote that the MJC “fail[ed] to provide a rigorous, balanced, and objective assessment” of the system because it used “inappropriate data” and relied on whether police reports were filed in the incident. 

But other studies, too, show varying levels of efficacy for the technology. In 2021, Chicago’s Office of the Inspector General concluded that only 9.1% of gun-related criminal offenses were found when the Chicago Police Department responded to ShotSpotter alerts. 

In Cambridge, former Police Commissioner Branville Bard wrote a memo during a city council probe of the system in 2021 giving data that showed that over six years, only 35 out of 105 activations were confirmed as a shooting incident — a false positive rate of over 65%. 

During that time, 13 arrests were attributed to the system, which cost the city $50,000 a year, according to the Cambridge Public Safety Committee. 

Based on records obtained from Cambridge by Boston.com, ShotSpotter has only gone off in the city five times during the past year. Three records were withheld due to active investigations into the incidents, but two others, one in November and one in April, were found to be false alarms likely triggered by fireworks. 

In Boston, public records requested by Boston.com showed that out of 571 activations of the system in the past year, about 250 were marked as “shots fired” or “multiple shots fired.”

Apart from the system’s accuracy, studies have shown that the system has little, if any, impact on reducing firearm related crime. A study of ShotSpotter data between 1999 and 2016 by the Journal of Public Health found that implementing the system “has no significant impact on firearm-related homicides or arrest outcomes.”

“Policy solutions may present a more cost-effective measure to reduce urban firearm violence,” the authors write in the study’s abstract. 

“Blind faith in ShotSpotter”: Critics raise concerns of racism, flaws in the system

Kade Crockford, the director of ACLU Massachusetts Technology for Liberty Program, said some of the major concerns around the system come from over-policing of minority and underprivileged communities — where critics say most of the sensors are installed — and even false arrests. 

In 2020, Michael Williams, a Black man, was arrested in Chicago under suspicion of shooting a passenger in his car. Prosecutors said ShotSpotter audio indicated Williams shot and killed the man, with little evidence besides the audio and nearby security video footage. 

“The arrest was baseless… Officers put blind faith in ShotSpotter evidence they knew or should have known was unreliable in order to falsely arrest and prosecute Mr. Williams for murder,” a lawsuit filed by Williams against the City of Chicago read. 

Williams spent 11 months in jail before the charges against him were dismissed in 2022 based on a lack of evidence. Crockford said the incident raised concerns about overzealous police officers, mainly getting alerts in minority neighborhoods, arresting suspects based on ShotSpotter activations. 

Another incident in Chicago involved officers firing at a 14 or 15 year old child when they received a ShotSpotter alert which turned out to be fireworks. In 2021, 13-year-old Adam Toledo was shot to death by police responding to a ShotSpotter activation.

The questionable accuracy of the system but the high-tension situations police are expecting when responding to an alert can be a dangerous mix, Naples-Mitchell said. 

“Police are being sent assuming there’s a shooter,” she said. “Police are sent into neighborhoods looking for evidence of a shooting which has them on extremely high alert and can make for very dangerous encounters for members of the public and for police, frankly, if they’re expecting there to be someone armed and dangerous when there is no one.”

Incidents like these, Crockford worries, only heighten mistrust between communities and police — the opposite of what ShotSpotter aims to help police do.

“If these kinds of things are only happening in Black and brown neighborhoods, and they’re most likely happening to Black and brown people, it contributes to a sense of mistrust, which undermines what the police continue to say they really need, which is public participation in criminal investigations into things like murders and shootings,” Crockford said. 

Though the exact location of ShotSpotter devices is confidential, leaked data analyzed by WIRED earlier this year showed that nearly 70% of people who live in a neighborhood with a ShotSpotter sensor identify as Black or Latine. 

According to the data, most sensors in Boston are concentrated in Roxbury and Dorchester, some of the most racially diverse neighborhoods in the city.

Naples-Mitchell added that feelings of constant surveillance and increased police contact in communities of color correlate to higher rates of physical and mental illnesses like hypertension. 

“These are significant quality of life and life outcomes that affect mortality that may also be related to having this very stressful set of encounters and experiences shaping one’s neighborhood,” Naples-Mitchell said.

ShotSpotter has also faced suspicion from critics about whether microphones can pick up street-level conversations. 

Crockford pointed to a 2011 incident when the sensors recorded a street argument accompanied by a fatal shooting in New Bedford. The system has also picked up voices in Oakland, Calif., according to local reports. 

The company fervently denies that the microphones have voice recording capacity. In response to questions from Boston.com, SmartThinking pointed to a study from the New York University Policing Project that concluded that “the risk of voice surveillance was extremely low in practice.” Chittum also said the microphones on the sensors “are not specialized in any way” and purge audio every 30 hours. 

More broadly, Crockford doesn’t believe this method of policing is helping meaningfully reduce crime. And when it comes to people not dialing 911 when they hear gunshots, that itself is an indication of a lack of trust in police, they added. 

“It’s adopting a technology to get around that trust issue, instead of police trying to address it,” Crockford said. 

“Force multiplier”: Police say ShotSpotter alerts save lives

According to the Brookings Institution, only about 12% of gunfire incidents lead to a 911 call. This is one of the primary selling points SoundThinking uses for ShotSpotter — it allows police to get to the scene of a crime faster, helping potential victims get medical assistance and allowing officials to collect ballistic evidence. 

Worcester Police Chief Paul Saucier says the system allows the city’s police department to build trust with the community by showing up to crime scenes even when nobody dials 911. 

“We try to look at ShotSpotter as being a community policing program. If nobody’s coming to your community when there’s firearms being discharged, that’s a problem,” Saucier said. 

Saucier told Boston.com that in 2023, Worcester had 116 shooting incidents, 70 of which involved ShotSpotter. Forty-one of the 70 did not involve a call to police. 

In response to claims from critics that the sensors are used mostly in minority communities, both Chittum and Saucier pointed to the fact that sensors are installed based on historic crime data. 

“Gun violence is not spread evenly across the country. It disproportionately affects many communities of color,” Chittum said. “I think ShotSpotter is an essential part of the public safety infrastructure for protecting communities like that.”

Saucier called the system a “force multiplier” for Worcester’s Police Department, which has used the system since 2014 and employs sensors in an eight mile stretch of the city, by allowing officers to get to the scene quicker.

“You have to put it in areas where the most gunshots and homicides and victims of shootings occur. You’re not going to put it in an area that’s never had a shooting,” he said. “For the most part, people in [high crime areas] want to see the police.”

Chittum acknowledged that the technology does occasionally miss shots fired or flags non-shooting related sounds. 

“Like any technology, it’s not perfect. We don’t say that it is 100% accurate because it can’t be,” he said. 

As for police responses to ShotSpotter alerts, Chittum said it’s up to departments to analyze and correct how they respond to incidents, and that ShotSpotter is just a tool to help law enforcement.

“It’s up to our customers, the police, to respond and investigate how they use this tool — [that is] an important part of its effectiveness,” he said. 

Boston and Somerville leaders, Massachusetts legislators have concerns about the system

Chicago, one of the cities most studied and criticized for its use of the technology, will end its contract with SoundThinking in September at the behest of Mayor Brandon Johnson. He said the city spends $9 million a year on the technology, which he called “unreliable and overly susceptible to human error.”

The Fall River Police Department stopped using the system in 2017 after four years because it had a 41% error rate in the city, The Fall River Herald News reported at the time. The system cost the city about $90,000 a year, according to The Herald News. 

In March, Somerville Councilor at Large Willie Burnley Jr. introduced an order to terminate the city’s contract with SoundThinking, saying the sensors are “always listening” for gunfire-like sounds.

Burnley also cited leaked information of ShotSpotter locations, which are reportedly installed in minority-populated areas like East Somerville and Prospect Hill. 

“The poorer and more diverse folks of East Somerville, living in neighborhoods with constant high levels of traffic noise, including cars backfiring, get constant audio monitoring from ‘35 sensors’ in case there is a ‘gunshot-like sound.’ The richer folks of West Somerville don’t,” Burnley wrote. 

Earlier this month, Senators Elizabeth Warren and Ed Markey, and Representative Ayanna Pressley sent a letter to the Department of Homeland Security urging the office to investigate federal funding of ShotSpotter, saying it may violate the Civil Rights Act of 1964. The Act prohibits recipients of federal financial aid from discriminating based on race, color, and nationality. 

The letter points to leaked ShotSpotter locations that show sensors to be concentrated in minority communities. Legislators also say the system is inaccurate. 

“Studies have also shown that biased deployment of the system can perpetuate the over-policing and unjustified surveillance of communities of color, exposing residents to police interrogations, confrontation, and potentially creating dangerous situations for residents,” the letter reads. 

“I’m concerned that this flawed, ineffective technology subsidized by the federal government is exacerbating racial bias in policing and could violate civil rights law,” said Warren in a statement to Boston.com. “This tool is disproportionately deployed in Black and Brown communities, which raises alarms about unjustified surveillance and over-policing in communities of color. “

But Chittum said critics point to the wrong statistics when criticizing the technology. 

“[Critics] will point to maybe half a dozen cities that have canceled their contract or not renewed over the course of over a decade, instead of pointing to the 170-ish that have had the system renewed or have expanded the system,” he said. 

When asked if there is evidence that ShotSpotter reduces crime rates, Chittum admitted that “ShotSpotter alone is not a solution.”

“There is no single solution to gun violence in America,” he said. 

Crockford, though, believes the  money going toward ShotSpotter can be implemented in better ways. 

“We’ve seen zero evidence of any public safety benefit, and we have seen real civil rights harms,” they said. “Public officials… really ought to be asking themselves, is there something that we could be doing with this money that would not cause these civil rights or civil liberties harm, and would instead help us address the underlying problem of mistrust within the community?”