Cruz was primarily talking about the improved response time the automated gunshot detection system affords police, a reported 63% quicker than responses to 911 calls. This would be a bigger deal if Pittsburgh police didn’t already have the fastest responses of any major city in the country, per a study by Matrix Consulting Group commissioned by city council last year. The controller’s audit shows that even in 2022, the slowest year they examined, police were en route to the scene within seven minutes on average even for gunshot reports obtained through 911 calls, an outstanding performance. Nonetheless, ShotSpotter alerts lead to arrests less than one percent of the time, and have not been correlated with any decrease in shootings.
Per the audit, the city will have spent about 8.5 million dollars in total on ShotSpotter by the end of the year, money that would have been much better used on affordable housing, ambulance repair, or a host of other projects. Part of the reason police are so attached to the devices is that calls to 911 about gunshots are in steep decline, and they want ShotSpotter to fill in the gap. A chart in the audit shows a steady reduction in gunshot-related 911 calls starting in 2015, leveling off during the pandemic years, and resuming in 2022. 2024 numbers are about a third of what they were in 2012, despite shootings being essentially constant over that period.
This raises a fascinating question. What on earth is going on here? Is the decline in gunshot-related calls reflected in all 911 calls? Are people who are aware of ShotSpotter less likely to call the police over loud bangs because they think the technology will do it for them? From the chart we can see the decline began in 2015, when ShotSpotter deployment was limited to a three-mile square block of Zone 5. By 2018, when the program was expanded to its current footprint, calls to 911 about gunshots had dwindled from around 1900 to around 1200 (it’s a little hard to tell from the chart, and we don’t get the raw data). For faith in ShotSpotter to explain the drop, Pittsburgh residents would have to have believed the sensors were far more widely distributed than they really were.
One possibility is not raised by the controller’s audit. 2015 was the first full calendar year after the acquittal of police officer Darren Wilson for killing Mike Brown in Ferguson, Missouri, the murder that started the Black Lives Matter movement. Confidence in law enforcement took a hit that it never recovered from, and the subsequent police murders of Breonna Taylor and George Floyd, among countless others, only accelerated the trend. It may be that fewer people call 911 because fewer people trust cops. Heisler’s report has no data that would let us test this theory, but it’s one worth keeping in mind.
Finally, we want to ask what ShotSpotter means for Pittsburgh’s Black and poor underclass. Would the main targets of police repression be better off if the technology was abandoned? Per the audit, ShotSpotter alerts lead to an arrest 0.44% of the time, compared to 0.33% for 911 calls. This is a minuscule difference that might not even be statistically significant (Heisler’s team didn’t check). ShotSpotter produces false alarms at a 25.23% rate, slightly over a percentage point more than 911. Again, this slight difference is probably not important.
What does appear significant is the number of alerts produced by ShotSpotter. A chart in the report shows that police responded to nearly 2,000 of the automated alerts in 2024, each requiring multiple officers per department policy. This reduces the so-called discretionary time in which cops have nothing to do but cruise the streets looking for homeless people to harass. The Matrix report linked above shows that Pittsburgh police already enjoy more discretionary time than any other department in the US.
What about the expense? Pittsburgh’s contract with SoundThinking, Inc., ShotSpotter’s vendor, is not available on the city contracts portal, but the 8.5 million since 2014 noted in the audit could have paid for any number of services that are actually useful. It could also have gone toward harmful measures such as police recruiting or more surveillance cameras. Even if ShotSpotter is eliminated, its budget is likely to be retained by the police. We might be better off keeping it and letting it waste the cops’ time.