A Nashville school district invested about $1 million in AI gun identification software, the school district said, leaving some to wonder what went wrong in detecting a school shooter in the halls.
More than a week ago, 17-year-old Solomon Henderson carried out a shooting inside Antioch High School in suburban Nashville, killing a classmate and wounding another. Henderson died after turning the gun on himself, police said.
The gun detection system at the school failed to detect Henderson’s weapon during the January 22 shooting because he was too far away from the surveillance cameras used to identify a weapon and activate an alarm, according to district officials. Meanwhile, the system’s CEO says the gun could not be detected because it was not visible.
The gun detection technology, called Omnilert, was deployed at all the schools in the district in February 2024 to bolster security, a spokesperson for Metro Nashville Public Schools told CNN.
Using AI technology, the system can identify a gun threat before shots are fired so local law enforcement can respond more quickly, and purportedly, prevent a tragedy from happening, according to the company’s website. But it’s only one tool that can be used in conjunction with other measures.
Omnilert’s software is relatively new and doesn’t work perfectly all the time, according to the company’s CEO, Dave Fraser, who earlier expressed his condolences to the school and community. And while there is no harm in implementing the technology if a school board can spare the cash, gun safety and surveillance experts say there is no concrete evidence gun detection software is effective at preventing school shootings.
Here’s how the technology works, what experts say went wrong and whether prevention is possible.
How does the technology work?
As the constant threat of mass shootings persists, school districts across the United States have invested countless dollars into safety measures to keep would-be shooters out of the classroom – a crime prevention technique known as target hardening.
Gun detection software, like Omnitech, is the latest target hardening resource to hit the market, said Chad Marlow, a senior policy counsel at the American Civil Liberties Union who focuses on privacy and school surveillance technology issues.
“There’s no intervention that is guaranteed to prevent a school shooting. If so, we’d all be using it,” Marlow told CNN. “But there are some that may have a positive benefit.”
At most schools, surveillance cameras are used to keep an eye on any possible threats, according to Fraser. While helpful, these cameras are only useful after a tragedy has occurred and footage is used to work out how the incident unfolded, he added.
Comparatively, the AI technology used in Omnilert’s Gun Detect software is trained to watch surveillance video for weapons in real time, Fraser told CNN in an emailed statement. These cameras are monitored by AI technology 24 hours a day, seven days a week, he said. CNN asked the company how many school districts are using the product but has not heard back.
“If a gun becomes visible it is detected in less than a second and the detection information – a picture and a short video of the person with the potential gun, the location of the camera and the building – is sent to a human to make a verification that this is a real threat,” Fraser told CNN by email.
In most cases, the technology initiates an emergency response in less than 20 seconds after a weapon is detected, Fraser said. Ideally, authorities get to the school before shots are fired, he added.
“At the very least it provides actionable information for staff and law enforcement to help them react to the situation, including knowing the exact location of the assailant,” Fraser added.
