Startups are using AI to help stop mass shootings, hoping to stop another Nashville

 

On the heels of the 128th mass shooting in the US for 2023, the question again arises: How can these terrible shootings be stopped?

One potential legislative approach comes from President Biden, who again urged Congress Monday to pass an assault weapons ban immediately following the Nashville shooting at an elementary school where three children and three adults were killed. Police said the shooter was a woman armed with at least two assault-style rifles and a handgun.

In past shootings, new approaches to technology, including AI, have been forwarded as potential remedies.  The same applies to Monday’s shooting, but technologies are advancing apace, meaning last year’s AI approach could be soon replaced by something newer and potentially more effective.

Some history: AI has been used for several years by companies, mainly startups, to help security teams identify someone carrying a gun or deadly weapon by training a system with a myriad number of images of handguns, assault rifles and other weapons from all types of angles and in various lighting.  Inference algorithms could then help spot people carrying weapons to theoretically alert police to intervene before an attack took place.

Questions have been raised by privacy advocates over whether such systems are accurate enough to avoid falsely accusing someone legally allowed to carry a weapon. (The majority of US states allow people to openly carry handguns.)  Or, could the AI be picking out suspicious-looking people who might or might not be carrying a weapon, potentially raising questions of racial profiling?

As far back as  2019, Hexwave technology from Liberty Defense Holdings  was tested at Virginia’s historic capitol in Richmond, Va.  The system relied on 3D radar imaging technology with AI to detect metallic and non-metallic firearms, knives, explosives, suicide vests and other threats. 

In order to be effective, Hexwave requires a person to walk through a screening portal and works with mmwave technology and video analytics technology from MIT, CEO Bill Frain told Fierce. The technology is similar to how some passengers are screened at airports by raising their hands above their heads, he said.

“It checks for metallic and nonmetallic threats, including ghost guns and 3D printed guns,” he said. “We don’t check bags, only what’s on the person. It could be part of the security solution where multiple technologies can be used along with ours. That’s our goal, to prevent something like this shooting in Nashville from happening.” 

Liberty Defense was started in 2018 and has 28 employees and is just beginning to commercialize its products after several beta trials, he said.

More recently, a California startup called Gunsens began offering a “shooting threat detection platform” in 2022 that features a few barebones descriptions and photos on its website of how its product works.   It claims to alert first responders and the public “before shots are fired, savings lives.” By comparison, Gunsens says its competition sends alerts after shots are fired, “potentially resulting in loss of life.”

A company official told Fierce Electronics that Gunsens’ sensing technology is proprietary, but its web site does reveal it can be installed anywhere, similar to a wall switch to dim lights. The site says the proprietary solution “uses AI instead of video surveillance.” The company appeals to saving lives in classrooms as well as offices, stores and homes.

“We don’t know much about how much the perpetrator in Nashville approached the building or traversed the buildings, but our solution works indoors and outdoors, 24/7 and in daytime and nighttime,” said Gunsens Chief Revenue Officer Charles Bret Martin in an interview.

“If we see a portion of a gun, we can notify instantaneously of the threat,” he said. “Competitive tech can take 60 seconds to five to eight minutes, but ours is sub-one to two seconds.”

While Martin wouldn’t reveal what sensing approach is being used by Gensens, he did say “It will know the broad class of guns.”  He said sensing methods used by competitors such as magnetometers, thermal differentiation and video cameras “all have their place and limits.”

Gunsens is just beginning to commercialize its technology, Martin said, and is accepting orders. He refused to offer more details, however.

If the Gunsens’ edge over others is a near-instant capability to alert security, that could be an effective edge, but there are still many factors to consider when assessing threats.

If an AI system relies on cameras, how many cameras would be needed to cover an entire school campus? asked Jack Gold, an analyst at J. Gold Associates. “Yes, you can probably use AI to detect the shape of a gun and issue an alert, but that assumes you can see the gun in the first place. Concealed weapons won’t be seen.”

Also, if sensing is done with radar waves, it could work unless the crowd is too large or dense. “It’s more appropriate for one-off, while walking through a detector, than scanning a whole crowd,” Gold added. “And, if the sensor is fairly sophisticated, it’s not something you are going to deploy at every school in the country.”

As with many technologies, cost and effectiveness will inevitably drive adoption of AI-based approaches for better protection against gun attacks in schools, stores and offices.

The industry for AI security is still bound to grow, however, reaching $1.2 billion in 2031, double what it was in 2022, according to Future Market Insights. One study by the National center for Education Statistics found that 91% of public schools in the US had security cameras for monitoring the school in 2019-2020, up from 61% a decade earlier.

ABC News recently reported Scylla, an AI security company, works with existing surveillance systems, adding in AI capabilities.  The company is based in Austin, Texas, and has 300 such installations across 33 countries. Meanwhile, ZeroEyes uses AI to track live video feeds to send an alert to an internal control room where company or school employees determine if a situation poses a threat.

It is far too early to tell what AI approach will become the most successful or the most effective to help combat mass shootings, but there does not appear to be a shortage of companies working to offer up solutions.

"Humanity is never at a loss for doing stupid things, but technology hopefully puts some sanity into the situation," said Gunsens' Martin.

RELATED: Police at Virginia capitol will test threat detection tech