Incidents associés
In 2017, Chicago police began to slowly expand a vast network of microphones mounted atop lampposts and utility poles---the electronic ears of the controversial gunshot detection technology ShotSpotter. This coincided with the creation of newly established high-tech, district-level intelligence centers known as Strategic Decision Support Centers (SDSC), which were developed through a collaboration between the Chicago Police Department (CPD), the University of Chicago Crime Lab, and the federal Bureau of Justice Assistance.
If ShotSpotter and its sensors are the CPD's metaphorical ears, then SDSCs are its brain---the centers cull and aggregate the tremendous amount of information police gather on the 2.7 million people who call Chicago home, and share that information with the central intelligence center at the department's headquarters. Staffed by a mix of sworn officers and civilian crime analysts, these centers serve as command posts not only for officers responding in real-time to potential crimes, but also for police who use historical crime data to predict future incidents.
ShotSpotter was to be a core pillar of these state-of-the-art facilities and a remedy to the city's persistent gun violence in beleaguered Black and Brown communities, which is perhaps why an analyst at one SDSC took a particular interest in the technology.
For at least nine months, between October 2017 and July 2018, Scott DeDore tracked ShotSpotter's accuracy in identifying confirmed gunshots. DeDore regularly shared his findings with Chicago police and ShotSpotter, and even attempted to hone the tool's precision by working alongside the company to install additional sensors, documents obtained through public records requests show. Over the course of those nine months, according to the records, ShotSpotter correctly detected a gunshot in 63 of 135 instances in which a person was struck, an accuracy rate of about 47 percent.
One month after DeDore sent his last available report, then mayor Rahm Emanuel signed a new three-year, $33 million contract with ShotSpotter (the company has since rebranded as SoundThinking). It covered 12 police districts---100 square miles---and made Chicago the company's largest customer at the time.
These records represent a look into a small corner of Chicago's southwest side from more than half a decade ago. But they offer a unique window into ShotSpotter and its role in an increasingly surveilled city. And they came at a time when the city was reinventing its policing strategy. Six years later, Chicago is again at a crossroad, as a new mayoral administration "reimagines" public safety and mulls the fate of ShotSpotter when its contract expires in mid-February.
SoundThinking did not answer a list of emailed questions from the Reader. In a written statement, Mark Page, the company's senior vice president of field engineering and customer service, said ShotSpotter is used around the clock by more than 160 agencies across the country, "each with Service Level guarantees of 90% or more in terms of detection performance." Page claimed the technology "consistently meets or exceeds this performance criteria" and said that SoundThinking incurs financial penalties "in the event that we fail to meet our performance commitments."
Page also pointed to a report from consulting firm Edgeworth Economics that found ShotSpotter maintained a 97 percent accuracy rate over the past four years. That study, however, was commissioned by SoundThinking and only determined whether the system accurately identified a sound as a gunshot. Further, the study relied on data provided by ShotSpotter---which, itself, was reported to the company by police departments---rather than conduct its own testing of the technology.
Page said that most shootings aren't reported to police. "In the time that it has been deployed in Chicago, ShotSpotter has led police to locate hundreds of gunshot wound victims where there was no corresponding call to 911. Those are victims who would not have received aid but for ShotSpotter."
The CPD did not respond to a request for comment.
DeDore began as an analyst with the CPD in June 2017, according to his LinkedIn. He was assigned to the tenth district's SDSC around the time the department was first unveiling the real-time crime centers and just a few months before ShotSpotter's expansion into the southwest side region.
As the gunshot detection technology went live in his district in October 2017, DeDore began to scrupulously document its ability to accurately alert officers to "bona fide" shootings in which a person was hit. According to records obtained from the CPD, DeDore tracked his findings in spreadsheets that listed, for each incident: the date, time, and location of the shooting; whether notification of the incident came from 911 dispatchers at the city's Office of Emergency Management and Communications (OEMC) or from ShotSpotter; the police beat in which the incident occurred; any evidence of a shooting; and whether the technology correctly alerted officers.
The technology's accuracy varied greatly over the nine months of available data. In the period between March 19, 2018, and April 15, 2018, for example, ShotSpotter alerted police to nine of the ten instances where a person was shot in the tenth district. But it didn't detect any of the six shootings between May 22, 2018, and June 11, 2018.
ShotSpotter's accuracy using the data and time periods from DeDore's analysis\ Source: Chicago Police DepartmentCredit: Amber Huff for Chicago Reader
In one of his first reports, on November 2, just a few weeks after ShotSpotter launched in the tenth district, he noted "13 MISSES with ground Evidence that occurred in the North End" of the district between October 16 and 30, 2017.
Further, DeDore found that from December 11, 2017 to February 11, 2018, ShotSpotter alerted the CPD to 65 percent of incidents when a person was shot in the tenth district. From February 12, 2018 to March 18, 2018, ShotSpotter alerted police to nine of the 17 incidents when a person was shot, a 53 percent accuracy rate.
It's possible some of the shots occurred indoors and were, therefore, undetectable by ShotSpotter. When the CPD expanded ShotSpotter into the tenth district, the city's then agreement with the company, which the *Reader *obtained from the Public Buildings Commission, stipulated the "system will be designed to detect at least 80% of the unsuppressed outdoor gunfire, with location accuracy to the shooter's location within 25 meters in at least 80% of the incidents."
DeDore regularly shared his findings in emails with dozens of police across multiple units within the CPD, including with members of the department's brass. Recipients included dozens of sergeants and lieutenants, some in the district, others in units including the narcotics, internal affairs, gang enforcement, and education and training divisions, documents show.
These emails included PowerPoint presentations with data tables that tracked how ShotSpotter's accuracy changed over time and maps that plotted the locations of undetected gunshots in the tenth district.
"I don't know if everybody in the new SDSC rooms appreciates the importance of improving ShotSpotter. This can only be accomplished by accurately documenting their MISSES," DeDore wrote in a February 22, 2018 email. "The city is paying for this it's a phenomenal tool, let's make it better!"
Documents also show DeDore shared his findings with various ShotSpotter representatives, and even worked with the company to increase the technology's accuracy. On December 5, 2017, DeDore sent an email to Doris Cohen, at the time a customer success and training manager for ShotSpotter, according to her LinkedIn. (Cohen is now SoundThinking's analytics director.) His email included an attached presentation that detailed an apparent conflict between the Latin Kings and Satan Disciples street gangs that had occurred a day earlier.
"On 4 December 2017, there were three conflict related shots fired incidents in the 010th district with no hits," the presentation notes. "ShotSpotter did not alert on any of the incidents, but did capture audio on each incident."
In another instance, on January 8, 2018, DeDore forwarded Cohen an email with an accuracy report for the period spanning December 11, 2017 to January 7, 2018. Nine minutes later, DeDore received a reply from Jeff Magee, ShotSpotter's customer success director until 2022, asking if the analyst was available for a meeting. DeDore responded with his availability, and Magee wrote back, "I will stop over Wednesday morning."
A couple days later, on January 10, DeDore emailed colleagues an updated ShotSpotter accuracy report. The presentation notes, "ShotSpotter has increased microphone placement in the 010th district and made adjustments in December based on geographic mapping of the documented 62 MISSES." In an email sent the following month, DeDore writes that "ShotSpotter made several trips here after I sent these reports. They then came and installed more microphones."
But efforts to improve the technology's accuracy in the tenth district over the course of the nine months of data proved relatively fruitless. Documents show the year-to-date accuracy rate remained below 50 percent for the entirety of the nine months of data.
It's unclear just how high up the command chain DeDore's analysis reached, or whether anyone in Emanuel's office knew of the accuracy reports ahead of the 2018 contract extension. But the documents show at least some within the department's leadership were aware of it. Then captain (now commander) Joseph Brennan and Commander James Sanchez were included in a March 20, 2018 email in which DeDore shared his analysis. DeDore, Brennan, and Sanchez appear to also have been invited to a meeting the same day---alongside other leadership like the then commanders of the ninth and 15th districts, Randall Darlin and Ernest Cato III---with ShotSpotter's Magee and Mike Will, who an email identifies as the company's vice president of service solutions. The meeting was about "Shotspotter Reported Misses and Service Enhancement," documents show.
"I will map all recent MISSES before then," DeDore wrote ahead of the meeting.
In July 2003, the CPD unveiled a pilot program that would install 30 remote controlled, 360-degree surveillance cameras on light posts in "Chicago's most violence prone communities," the department website notes, with the goal of "disrupting illegal narcotics operations." Police credited these so-called police observation device (POD) cameras with cutting drug crimes by three-quarters during the summer of 2003.
By September, Mayor Richard M. Daley would take the program citywide. As the number of POD cameras grew---first from 30 to 80, eventually to thousands across the city---so, too, did their supposed capabilities. This phase marked the city's first foray into gunshot detection. Second-generation POD cameras were equipped with technology from Tuscon, Arizona--based surveillance company Safety Dynamics that the CPD says could wirelessly transmit gunshot alerts to the OEMC, field control devices, district stations, and police headquarters.
Chicago's first deal with ShotSpotter came in 2007. But it, along with the pilot from Safety Dynamics, were short lived. A Chicago police spokesperson told NBC Chicago in 2010 that the department dumped both programs because they were costly and "not entirely effective."
The city gave ShotSpotter another go in 2012, this time reportedly financed with funds seized from civilians through a process known as civil asset forfeiture. Initially installed in just two of the more than 20 police districts---covering three square miles---ShotSpotter would soon wrap its tendrils around large swaths of Chicago. By September 2016, the technology's footprint had grown to 13.5 square miles, after the Public Buildings Commission approved nearly $1 million in funding to cover the entire seventh and 13th districts---Englewood and Garfield Park---in sensors.
ShotSpotter's growing coverage area accompanied the rollout of the department's first two SDSCs, in the same districts, at the start of 2017. The investment in its real-time crime centers marked a shift in strategy for the CPD that centered on "inserting technologies into district-level centers," according to a 2019 Rand Corporation report. The study notes Chicago police desired to "better integrate technology into CPD's policies and practices," leverage the "very large amount of information" it already maintained on Chicagoans, and address the real and persistent problem of gun violence.
SDSCs are small, converted conference rooms in district stations. They're staffed by a team of two to four officers and civilian analysts per shift, and supervised by a sergeant or lieutenant. These centers include access to a bevy of technologies, including ShotSpotter, POD cameras, geospatial predictive policing software, and social network analysis, that police heralded as cutting-edge public safety tools.
In spring 2017, the CPD opened SDSCs in four additional districts, including DeDore's tenth. By March of the following year, 13 districts would have SDSCs, with plans to eventually open one in every district.
The expansion of SDSCs and the introduction of new surveillance technology came at a time when the CPD faced a severe crisis of legitimacy. Decades of brutality, dismal clearance rates, the killings of Rekia Boyd and Laquan McDonald, and a federal investigation into a pattern of unconstitutional policing laid bare systemic issues within the department. An emphasis on high-tech policing, then, offered police an opportunity to rebrand.
City leaders from Emanuel to former Chicago police superintendent Eddie Johnson heralded ShotSpotter as part of a new wave of "data driven enforcement" that was central to bringing public safety to Chicago's long divested and heavily segregated south and west sides. In announcing the opening of the tenth district's SDSC in October 2017, Emanuel said the new focus on technology "is part of CPD's commitment to a crime strategy that has made us smarter in our deployments, faster in our response times, and more proactive in our community engagement."
At the same time, however, a member of that police department collected nine months of data that showed, at least in one district, in cases where a person was hit by gunfire, the technology missed as many shootings as it identified. None of the records provided to the Reader show police questioned the continued use of the technology. Instead, the department seemed resolute to maintain its relationship with the company.
Emanuel signed off on a three-year extension with ShotSpotter in August 2018. Johnson, in a press release from the company announcing the deal, made clear he thought the technology was here to stay. "ShotSpotter has been---and will continue to be---a key component in helping our agency rebuild trust with the communities that we are serving," the superintendent said.
Chicago's current ShotSpotter contract expires on February 16---but, even as the deadline looms on the horizon, Mayor Brandon Johnson has remained tight-lipped about its prospects, despite committing to end the city's contract with the company as a candidate. Since taking office, his administration has sidestepped questions about the city's future with the technology. He and other city officials have instead suggested that they are considering a range of options, but have so far declined to discuss specifics.
In June 2023, Chicago approved a $10 million payment to SoundThinking, according to the Chicago Sun-Times. Jason Lee, a top Johnson adviser, said at the time the payment was erroneously processed after former mayor Lori Lightfoot signed off on an extension through February 2024. South Side Weekly reported in November, however, that SoundThinking CEO Ralph Clark told investors Johnson "included funding for the continued use of acoustic gunshot detection technology" in his 2024 budget proposal, which the city council approved later that month. The budget includes a roughly $9 million line item for "Software Maintenance and Licensing." That same expense also appeared in the 2023 budget.
Ronnie Reese, the mayor's press secretary, repeatedly declined to make anyone available for an interview and would not comment on whether Johnson will end Chicago's use of ShotSpotter when the contract expires or how the city plans to spend the $9 million allocated for it. When asked by the Reader if he could provide any information on the impending decision, Reese replied, "No updates."
Meanwhile, more than 80 community groups and more than 500 people have signed an open letter calling on the Johnson administration to end the city's relationship with ShotSpotter. "More than just being a waste of city resources," the groups write, "the use of ShotSpotter results in deployments that are dangerous for Black and brown community members who historically have been killed, tortured, disappeared, and terrorized by CPD officers."
The letter calls for greater investment into areas like housing, education, and health care. "Every penny invested in surveillance tools demonstrates a lack of confidence in community-led solutions to safety," it says.
Whether the mayor stays true to his promise remains to be seen. But his top cop, Larry Snelling, might offer some insight. At a January community forum in Chinatown, Snelling, a longtime supporter of the gunshot detection technology, offered a full-throated defense of ShotSpotter and suggested that people who oppose it misunderstand how it works. "You tie sound detection to a lot of things it's not tied to. You can read studies---and I can show you how it works," the superintendent said. "Until you actually see it, you don't understand."
"Shot detection does one simple thing: it detects gunshots and it allows officers to get to those locations quicker."