By Nick Corn IV

 

As technology has advanced, law enforcement has adapted the use of new tools to attempt to reduce response times and keep emergencies from slipping through the cracks. ShotSpotter produces one such tool. ShotSpotter bills itself as a “precision policing platform [which] helps local, state and federal law enforcement respond to, investigate and deter crime.”[1] ShotSpotter uses a network of microphones strategically placed across the patrolled area to listen for loud bangs. [2] Once a bang is detected, a computer algorithm makes a determination on whether that bang was a gunshot or not.[3] A human analyst then reviews that algorithm’s determination and decides if it is appropriate to dispatch police.[4]

ShotSpotter claims that its system has resulted in an increase in reporting gunshots from an estimated 12% of the time before the use of their technology to 90% of the time after.[5] However, ShotSpotter can prove to be quite expensive, costing between $65,000-$95,000 per square mile every year.[6] In an effort to get the most out of their money, police departments have taken to strategically placing the microphones in densely populated inner-city areas, which often happen to be communities of color.[7] So, what happens when the new tools used to detect gunshots can prove to be unreliable?

ShotSpotter claims that their technology has a 97% aggregate accuracy rate with a false positive rate of less than half of one percent.[8] However, ShotSpotter has never been peer-reviewed by academics or experts.[9] That said, a report done by the Office of the Inspector General for the City of Chicago found that between January 1, 2020, and May 31, 2021, that of the 41,830 dispositions of police to a scene as a result of ShotSpotter indicating a gunshot, only 4,556 resulted in evidence of a gun-related offense being found.[10] Even presuming that some evidence of gun crimes may be removed or concealed prior to law enforcement arriving on the scene, this number still suggests that the accuracy rate purported by ShotSpotter may be far too optimistic in their own capabilities. Similar studies done by cities on false alerts, such as Charlotte, North Carolina, and San Antonio, Texas, have led them to end their contracts with ShotSpotter.[11] However, this technology is still being used in 117 cities in 34 states and U.S. territories.[12]

ShotSpotter data can present difficult evidentiary situations as well. While the algorithmic data can be used to corroborate evidence found at a crime scene to boost accuracy, some courts filed charges against Defendants in the absence of any physical evidence, such as a gun, based on ShotSpotter data. [13] One such instance is Michael Williams, a 65-year-old Chicago man who was charged with the first-degree murder.[14] Williams was accused of shooting and killing a 25-year-old man who Williams says he picked up on the way back from making a cigarette run.[15] Williams claims that a car pulled up beside him at an intersection and shot into his car, killing his passenger, while prosecutors claimed Williams shot his passenger from inside the car.[16] Williams was charged despite nothing in the police report citing any motive, a total lack of eye-witnesses, and no gun found at the scene.[17] Williams was charged mostly due to the data collected by ShotSpotter, despite the fact that it originally identified the sound as a firecracker, with 98% confidence, at a location about a mile away from where Williams and his passenger were at the time before the data was re-labeled by ShotSpotter employees.[18] The charges against Williams were eventually dropped at the request of prosecutors when they determined “the totality of the evidence was insufficient to meet [their] burden of proof.”[19]

Williams surely isn’t the only person who has faced this exact scenario. Tania Brief is an attorney at The Innocence Project, a legal non-profit who seeks exonerate innocent people who are incarcerated for crimes they did not commit.[20] Brief stated that “[t]he concern about ShotSpotter being used as direct evidence is that there are simply no studies out there to establish the validity or the reliability of the technology. Nothing.”[21] Understandably, organizations like The Innocence Project would be wary of technology that they believe could contribute to more innocent people being incarcerated due to unreliable direct evidence. ShotSpotter surely has advantages, such as a decreased dispatch time of less than 60 seconds as compared to an average of 4.5 minutes and a decrease in the transport time of gunshot victims to hospital from 10.3 minutes to 6.8 minutes.[22] However, at this point in time, this new tool of law enforcement may prove to be unprepared for reliability in the courtroom.

 

[1] See Gunshot Detection, ShotSpotter (Nov. 5, 2021), https://www.shotspotter.com/law-enforcement/gunshot-detection/.

[2] See James Clayton, Inside the Controversial US Gunshot-Detection Firm, BBC (Oct. 29, 2021), https://www.bbc.com/news/technology-59072745.

[3] See id.

[4] See id.

[5] See Gunshot Detection, supra note 1.

[6] See ShotSpotter, ShotSpotter Fact Sheet 1 (2016), https://www.shotspotter.com/system/content-uploads/ShotSpotter_Fact_Sheet_-_final_draft_12.13.pdf.

[7] See Clayton, supra note 2.

[8] See ShotSpotter, ShotSpotter Respond Q&A 2 (2020), https://www.shotspotter.com/wp-content/uploads/2020/12/ShotSpotter-Respond-FAQ-Dec-2020.pdf.

[9] See Burke et al., How AI-Powered Tech Landed Man in Jail with Scant Evidence, AP News (Aug. 20, 2021), https://apnews.com/article/artificial-intelligence-algorithm-technology-police-crime-7e3345485aa668c97606d4b54f9b6220.

[10] See Furgeson & Witzburg, The Chicago Police Department’s Use of ShotSpotter Technology 14 (2021), https://igchicago.org/wp-content/uploads/2021/08/Chicago-Police-Departments-Use-of-ShotSpotter-Technology.pdf.

[11] See Burke et al., supra note 9.

[12] See Cities, ShotSpotter (Nov. 5, 2021), https://www.shotspotter.com/cities/.

[13] See Burke et al., supra note 9.

[14] See id.

[15] See id.

[16] See id.

[17] See id.

[18] See id.

[19] See id.

[20] See About, The Innocence Project (Nov. 5, 2021), https://innocenceproject.org/about/.

[21] See Burke et al., supra note 9.

[22] See Gunshot Detection, supra note 1.

Image Source: https://www.voiceofsandiego.org/topics/public-safety/shotspotter-sensors-send-sdpd-officers-to-false-alarms-more-often-than-advertised