The vast majority of shootings detected by a controversial police surveillance tool do not result in an arrest, but Detroit Police insist that ShotSpotter listening devices are driving down shootings this year. 

Deputy Police Chief Franklin Hayes and Stephen Lamoreaux, director of crime data analytics for DPD, updated the City Council’s Public Health and Safety committee Monday on ShotSpotter’s performance so far this year. The audio sensors were expanded to 40 square miles of Detroit in April after the council approved a $7 million contract with SoundThinking, Inc. in 2022. Hayes said ShotSpotter has had a “direct effect” on the reduction of shootings, though council members said police haven’t provided enough data. 

Related:

• ‘ShotSpotter’ says mistakes are rare, but Detroiters want more transparency
• Is ShotSpotter effective? Data on Detroit technology fuels debate

Only 2% of ShotSpotter incidents this year resulted in an arrest, and 3% of incidents resulted in a firearm being recovered. Detroit police made 332 arrests and recovered 585 firearms in 2023. 

There have been 13,725 reported gunfire incidents this year, as of Oct. 11, accounting for 48,371 individual rounds of gunfire. 

Council Member Gabriela Santiago-Romero said DPD needs to provide additional information to determine whether the technology has been effective. 

“This is a very expensive piece of technology,” Santiago-Romero said. “I would love to see it work and I would need to see that in numbers … The numbers right now aren’t screaming success because I don’t potentially know the full story of what is coming off the streets.”

Council Member Scott Benson said he’s an early supporter of ShotSpotter – the devices were tested in his northwest Detroit district, which paved the way for a multi-million dollar contract with ShotSpotter. However, he said DPD has not provided quantifiable metrics to determine whether the investment is worth the cost. 

“This is a financial resource that we could be placing someplace else; is it a good use of our money?,” Benson said. 

Lamoreaux pointed to two major benefits of the technology. He said ShotSpotter helps police recover shell casings and other evidence that can be used to track crime guns. It also alerts police to shootings they otherwise wouldn’t have known about, he said. 

“There are a lot more leads provided by ShotSpotter,” he said. “It looks like people either have accepted that shots fired are a fact of life in their neighborhoods or (they) do not want to report (gunshots) out of fear of potential retaliation.” 

Lamoreaux said a previous analysis found very few shootings identified by ShotSpotter were also reported by residents. Lamoreaux said residents called 911 in under 20 percent of ShotSpotter alerts. Hayes told council members that the technology is a deterrent that can also “ultimately change the behavior” of shooters.

Benson said the data shows a small percentage of people commit a majority of violent crimes, so even a handful of arrests could have a major impact. 

“We need to look at how we’re measuring these metrics as well and be really creative about how this is working,” Benson said. “The chair (Santiago-Romero) is right, this is very expensive. Let’s make sure we’re doing our return on investment metrics properly. 

The network of audio sensors are supposed to triangulate the location of gunfire sounds within an 82-foot-radius. Gunfire sounds are reviewed by SoundThinking analysts who send an alert to the Detroit Police Department. Lamoreaux said the entire process can take under one minute. 

Nonfatal shootings and homicides are down overall this year – including in areas where ShotSpotter isn’t active – dropping by 9 percent and 11 percent, respectively. Hayes said there are 111 fewer shooting victims in Detroit so far this year. 

Nonfatal shootings are down in six of nine police precincts using ShotStopper. Shootings are up by 53% in the 2nd Precinct and 16% in the 6th Precinct, with no change recorded in the 10th Precinct. 

ShotSpotter devices only cover a portion of police precincts they are deployed in. Lamoreaux said shootings dropped faster near precincts where ShotSpotter is not deployed, though the difference is “probably statistically insignificant.” He did not have the data available on Monday. 

Detroit started using ShotSpotter in March 2021 with 6.5 square miles of coverage across the 8th and 9th police precincts. A $1.5 million contract was extended last year, and the City Council also narrowly authorized another $7 million contract to widen the network.

Four of nine council members who voted against the latest contract argued that DPD failed to prove the technology can reduce crime. 

The coverage area increased to 40 square miles this year, including nine out of 11 police precincts. ShotSpotter devices cover roughly one-fifth of Detroit. 

Detroiters have no way to know whether they are near a ShotSpotter device. Even Detroit Police Chief James White doesn’t know where the devices are located, Lamoreaux said. The sites were chosen by ShotSpotter, which also installed them. 

Benson said he wants to know whether the ShotSpotter alerts result in more arrests and guns recovered compared to shootings identified through a 911 call. 

All recovered bullet casings are put into a federal database known as the National and Integrated Ballistic Information Network (NIBIN). Lamoreaux said the database helps police identify when guns are linked to other crimes.

Last year, ShotSpotter alerts led to 624 NIBIN leads and identified 465 illegal firearms. Those statistics more than doubled since ShotSpotter expanded. There have been 1,628 NIBIN leads and 1,197 unique crime guns identified this year. 

“(ShotSpotter) provides an invaluable investigative lead about how a firearm is used in multiple incidents, allowing us to take leads from one incident and use it to investigate another,” Lamoreaux said. 

Benson also suggested more coordination with community violence intervention programs to provide social services to shooting victims and residents living in areas with high amounts of gunfire.