One of the major problems with enterprise-level security methods is that they rely on human interaction. For example, if a multi-level retail complex has cameras monitoring areas of the store, those cameras need to be watched by trained personnel who may or may not be paying attention to the right screen at the right time, when a crime is taking place.
Things become even more complicated when you have multiple locations being monitored by an external security company, such as stores that are open 24 hours a day, service stations and so on. These type of locations are vulnerable to robbery and rely on a rather drawn-out sequence of events taking place in order to initiate a response.
First, the criminal enters the store, dons a mask and pulls out a knife. They then approach the attendant and demand money. At this point – with the crime already in progress – the security company representative notices something unusual on a bank of cameras and looks closer. Once it’s obvious what is happening, they contact emergency services.
By utilizing artificial intelligence, a security company, Defendry, has found a way to speed up this process exponentially. Not only are security screens monitored by an inorganic system that can maintain focus on everything at once, but the AI can also detect criminal actions such as an attendant raising their arms, an unusually large item in the hand of a customer, or merchandise being placed in a pocket or bag without approaching the counter.
But that’s not the best part, the system can also watch for actions that may preempt a criminal act. These actions may include a balaclava being put on before entering the store, a car being parked illegally close to the door, masked people ion the forecourt of a service station, or someone attempting to avoid security cameras. This means that crimes can be predicted and action is taken faster than a human – increasing police response time and increasing overall security effectiveness.