SALT LAKE CITY: Artificial intelligence company, Banjo, based in Park City, has partnered with law enforcement agencies across the state. Their goal is to more quickly respond to emergencies and fight crime faster.
“Every second counts,” Banjo founder and CEO Damien Patton told KSL Newsradio’s Dave and Dujanovic, when it comes to emergencies.
And in his eyes, artificial intelligence (AI) can help.
Law enforcement agencies in Utah are using Banjo’s AI in cases of missing children, for example, and finding them “sometimes in seconds,” he said.
How artificial intelligence works?
Patton describes Banjo as a business that is “unlocking silos of data.”
Its artificial intelligence uses a system called an “event-detection engine” that scans a stream of data from a variety of sources.
Those sources include 911 calls, social media posts, traffic cameras, CCTV cameras, and more.
The scan begins when Banjo AI detects an anomaly from the stream of data the technology scans.
The sound of gunfire in videos posted to social media in an anomaly. So is a signal given off when an airbag deploys are both examples of an anomaly.
Banjo’s system deletes the personally identifiable information it collects. Patton says that ensures user data is protected.
“Then it goes into our op center, and our team is able to disseminate fact from fiction,” he said.
A team of humans at Banjo ensures that the system is not reporting false positives.
Banjo alerts authorities once an anomaly is detected and verified.
Could artificial intelligence make driving safer?
Banjo already partners with law enforcement and other government agencies, like the Utah Highway Patrol and the Utah Department of Transportation.
And through partnerships with car manufacturers, Banjo can detect airbag deployments, Patton told Dave & Dujanovic,
“When an airbag goes off in a car, that doesn’t mean that 911 is notified immediately,” Patton said.
“But with the power of Banjo, and the partnerships we have with car manufacturers, if your airbag does go off we’re able to alert highway patrol minutes faster than the average 911 call would come in.”
“Not only does it help save your life, and get you to the hospital faster, it also clears the accident faster and ensures there’s not a secondary accident,” he said.
Patton says that Banjo will meet with the Utah Highway Patrol after a wrong-way driver on I-215 in Holladay caused a crash that killed two people.
He said Banjo detected the wrong-way driver incident as it happened. He also says that Banjo is now talking with authorities about how to prevent similar incidents before they happen.
Detecting crime in real-time
Before partnering with law enforcement, Banjo worked with news agencies.
Their philosophy changed after detecting, in real-time, the 2017 shooting in Las Vegas at the Mandalay Bay resort.
Patton told Dave & Dujanovic in a previous interview that the Banjo system was able to zero in on the shooter’s location in the Mandalay Bay.
Banjo alerted news agencies across the country with that crucial information, but outlets took “almost an hour” to notify the public, he said.
“That was the tipping point for me, personally,” he said. “We can no longer wait around for others to save peoples’ lives.”
As a result, Banjo started working with public safety officials.
“They’re the ones that need the information,” Patton said.
He described a recent example of a missing child—one that had not been abducted, but law enforcement reacted as if the child had.
Banjo’s technology gave law enforcement information that helped locate the child “within the hour,” Patton told Dave & Dujanovic.
Privacy concerns of artificial intelligence
It doesn’t take long for the issue of privacy to arise when talking about artificial intelligence. How do companies like Banjo ensure that data won’t be misused?
Patton argued that leaders in the tech industry and in government should create so-called “bumper-rails” that are connected to the use of AI.
In the case of Banjo, a bumper-rail would be the act of stripping personally identifiable information out of the data they collect.
“Any social media content we receive is faceless and nameless,” Patton said.