Skip to content

ShotSpotter sped Durham's police response to gunfire by 1.2 minutes

In Durham's 12-month pilot, evaluators found that median police response time fell 1.2 minutes more in the ShotSpotter area than elsewhere. They could not show an effect on gun violence.

  • 2 for an organization

    ShotSpotter cut the median response time by 74 seconds against the rest of the city and brought police to 73 confirmed shootings that they knew about only through alerts.

  • Not rated society

    The evaluators could not determine an effect on gunshot injuries, and they call it plausible but uncertain that a faster response saved one life.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Reported benefit.
Evidence

foundstudiedreportedclaimed, the grade of this benefit

Only the evaluators support the benefit.

Size

No magnitude yet: distribution is not documented.

Compared with

The rest of Durham, before and after the pilot. Measured.

Without the system
406 seconds (6.8 minutes) if the target area had followed the control pattern
With it
332 seconds (5.5 minutes), the observed 2023 median
The system’s part
The evaluators compared the pilot area with the rest of the city before and after the pilot, which isolates the service's part in the faster response. The service includes staff who review each alert, and the report does not separate them from the sensors and classifier or say whether the classifier learns from data.
Learns from data

Shown. SoundThinking's blog says that ShotSpotter filters detections with a machine learning algorithm that was trained on 25 years of gunshot and non-gunshot incidents. The Duke report does not describe how the classifier works.

“The resulting image mosaic is analyzed by a machine learning algorithm trained with gunshot and non-gunshot incidents collected over 25 years of real data to filter out incidents that are clearly not gunshots.” Source

Not yet checked
LimitsCosts and harmsLater outcome
Benefit type
Reduced Task Time
System
ShotSpotter gunfire-detection service
Made by
SoundThinking, Inc.
Used by
Durham Police Department (City of Durham)
Where
Durham, North Carolina, United States
Stage
In a pilot
Date
, when the source was published
Kind
Aggregate result
Broad benefit
Physical health, Operational improvement (crime reduction or prevention)
Collections
Scrutiny and findings
Last revised
1 November 2026

Sources

  1. Evaluation report, by an evaluator that the makers or users commissioned.

    Cook, Philip J. and Adam Soliman. Evaluation of Durham's ShotSpotter Installation: Results of a 12-Month Pilot Project. Wilson Center for Science and Justice, Duke Law, February 2024. The Durham Police Department arranged the evaluation.

    Median response time during the pilot period declined by 1.2 minutes in the target area compared with the rest of the city
  2. Other, by the makers. Context or contrary evidence.

    SoundThinking. SoundThinking's Approach to Fourth of July Safety. Company blog, 29 June 2026.

    The resulting image mosaic is analyzed by a machine learning algorithm trained with gunshot and non-gunshot incidents collected over 25 years of real data to filter out incidents that are clearly not gunshots.

Details

AI system

Workflow

Acoustic sensors covered a three-square-mile area chosen for its high rate of gun violence. Staff at the SoundThinking Incident Review Center confirmed each alert, which reached the police dispatch system and the nearest patrol cars almost always within 60 seconds. Police treated every gunshot notification as a Priority 2 incident that sends two patrol cars.

Follow-up

Conditions for replication

The gunfire occurs outdoors, since the system is not designed to detect gunfire indoors or in vehicles. Officers treat an alert as a lead and weigh the totality of the circumstances before they detain anyone.

Entry details

Sector

Policing (gunfire detection).

Purpose

SoundThinking built ShotSpotter to detect gunfire with acoustic sensors and alert police to its location. The evaluators say that it is designed to make shootings known to police and to cut the time until an officer arrives.

30 more questions not yet checked

model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this benefit

Peter H. Vartanian. (2026). ShotSpotter sped Durham's police response to gunfire by 1.2 minutes (AIBD-1218). Artificial Intelligence Benefits Database, edition of 2 November 2026. https://petervartanian.xyz/aibd/records/AIBD-1218.html (archived at https://doi.org/10.7910/DVN/AADSII, version 9.0)

BibTeX Data Field definitions