A New Approach to Urban Gunshot Detection
Chicago is testing a new way to identify gun violence in real-time. A local startup called SplitSec has released an application designed to turn standard smartphones into acoustic sensors. This technology aims to alert nearby individuals when shots are fired. It seeks to replace the need for static, city-wide audio hardware.
Sri Deivasigamani started the company in 2025. The core idea came from witnessing incidents where people remained stationary for long periods during a shooting. Many residents simply did not know what they were hearing or where the danger originated. By using the microphones already present in most Android devices, the software tracks audio patterns to identify potential gunfire events instantly.
Technical Implementation and Community Trials
Modern smartphones possess significant computational capacity. SplitSec leans on this power to run an artificial intelligence model locally on the device itself. This design allows the application to stay dormant until it detects specific acoustic signatures associated with a firearm discharge. The process operates similarly to how voice-activated assistants wake up when they hear a specific wake word.
Currently, the company is conducting no-cost trials within the 21st and 42nd wards of Chicago. There are ongoing discussions regarding expansion into the 5th, 29th, and 38th wards. Users participating in these tests provide feedback on accuracy and performance. While early data showed some false positives, the company is adjusting its model to better distinguish between sudden, loud noises and actual gunfire.
Privacy, Reliability, and Future Scaling
Privacy remains a significant point of interest for the public. Deivasigamani states that audio analysis happens entirely on the local device. The company does not record or transmit audio unless a user reports an incident, at which point the clip is used solely to refine the machine learning model. The goal is to build trust through transparent operations while providing a safety tool that relies on user confirmation, similar to a crowd-sourced navigation platform.
This approach differs from the city’s previous reliance on external hardware like the SoundThinking system. That contract ended in 2024, leaving the Chicago Police Department without its former sensor network. SplitSec is not yet working with police, though the founder notes that sharing anonymized alert maps could assist authorities if such a partnership were to materialize. As the trials continue, the company must prove it can scale its user base effectively while maintaining high accuracy in diverse urban environments. The broader significance lies in whether personal consumer electronics can replace or augment municipal safety infrastructure.