A teenager injured in a January 2025 shooting at a Nashville, Tennessee high school recently sued the maker of an “AI gun detection” system that failed to detect the gun that killed two people, including the shooter.
AllegedlySecurity company Omnilert either knew or should have known “significant operational limitations in the gun detection system that could result in detection failures during actual emergencies, including camera placement, gun proximity to camera sensors, camera angle, lighting and gun visibility,” according to a filing filed in Davidson County court last month.
Omnilert co-founder Ara Bagdasarian declined Ars’ invitation to answer questions about the lawsuit. Another defendant in the lawsuit, System Integrations, which resold the Omnilert system, also did not respond to Ars’ request for comment.
Board of Metropolitan Nashville Public Schools (MNPS) in 2023 has been confirmed a contract valued at more than $1 million to install an artificial intelligence detection layer over its district-wide network of cameras and related security infrastructure.
MNPS spokesman Sean Braisted said at the press conference According to the shooter’s position in relation to the cameras after the January 2025 shooting, the footage was “not close enough to accurately read and trigger the alarm.”
The claim often refers to Omnilert’s marketing copy on its website (eg stored on the Internet Archive a few days before the shooting), claiming that the company oversold its capabilities:
Omnilert went on to present that its AI-powered visual weapon detection “could have mitigated or prevented the tragedy at Marjory Stoneman Douglas High School,” referring to one of the nation’s most devastating school shootings to convey that its product would have prevented similar tragedies by identifying threats earlier…
Omnilert made no mention of false alarms, false positives, or detection limits on its commercial website prior to filming.
Chris Smith, one of the plaintiff’s attorneys, told Ars that the use of specific situational conditions in which the detection system is effective is questionable.
