AI Scannning For Weapon At Sports Venues & Arenas

Artificial Intelligence (AI) is being used across Britain to help secure sites from sports arenas to churches and schools. Indeed, some of the world's biggest venues, including Manchester Arena, where an Islamist extremist suicide bomber detonated a shrapnel-laden homemade bomb in in May 2017, murdering 22 young concert goers, is now using the technology.

The technology is being used to scan for weapons, including guns, knives and explosives as people walk between standing panels. If a weapon is spotted, security standing by is alerted. But there are emerging concerns that it does not work well enough.

The company which operates the AI, Evolv, sells security arches that contain a combination of electronic surveillance and AI software to sift through signals and spot one gun-carrying individual among thousands of people filing into a concert. 

They claim they can detect all weapons, although according to reports, the AI may not be capable of detecting certain types of knives, bomb components and some other dangerous items. 

The current extremely high alert level for venues has prompted Evolv to blend AI software with radar to cut down on false alerts and keep the crowds flowing into a venue. without time-consuming delays. Fans are asked to keep items in like their phone, keys, wallet, and other personal items in their pockets, as they walk through an Evolv arch scanner.

In 2020 Evolv was asked if it could build arches into doorways and examine ways to widen the space between sensors to allow larger crowds to go through more quickly. This technology, Evolv Express, was used to screen VIPs at this year's Brit Awards at London's O2 centre.

Traditional metal detectors throw up a lot of alarms for innocent metallic objects, creating a bottleneck for ticket holders who just want to get to their seat in a stadium or concert hall. One single Evolv dual-lane unit can scan upwards of 3,500 people walking through per hour, a pace that is ten times faster than traditional metal detector. 

Evolv’s research shows that about every 90 out of 100 fans walk through its systems without setting off an alert for additional security screening.

Security teams can change the settings on Evolv’s machines through the company’s software, typically from an iPad, to increase or decrease the sensitivity levels for Evolv’s AI to identify something as being potentially dangerous. According to reports, it is this feature that might be causing problems. 

The technology should not just be looking for the shape of a gun as defined by the software, but also for small shards of metal packed into a confined space to create shrapnel around an explosive device as was tragically demonstrated in Manchester Arena bombing.

Evolv’s computer vision sensors are trained to detect threats, such as bombs or firearms, through scanning shapes, sizes and density of items as fans walk through its security system.

AI and machine learning enable the scanners to create unique "signatures" of weapons that differentiate them from items such as computers or keys, Evolv says, reducing manual checks and preventing long queues. "Metallic composition, shape, fragmentation - we have tens of thousands of these signatures, for all the weapons that are out there... all the guns, all the bombs and all the large tactical knives." Evolv's CEO Peter George said last year. 

For several years, independent security experts have expressed doubts about some of Evolv's claims.

Last year, Evolv gave permission to the US National Center for Spectator Sports Safety and Security (NCS4) to test its technology and NCS4's public report, gave Evolv a score of 2.84 out of three - many types of guns were detected 100% of the time. But it also produced a private report, obtained via a Freedom of Information request which has been reported by BBC News along with emails between Evolv and NCS4.

This report gave Evolv's ability to detect large knives a score of just 1.3 out of 3. In 24 walkthroughs, Evolv Express failed to detect large knives 42% of the time. "The system was incapable of detecting every knife on the sensitivity level observed during the exercise," the report says.

Evolv:      BBC:     IPVM Research:    SportsTechie:      BusinessTelegraph:     CyprusNews:    Techregister:

You Might Also Read: 

Artificial Intelligence & Its Impact On Cyber Security: