Undoubtedly, artificial intelligence (AI) is the buzzword dominating 2023. While AI is not a new concept and has been utilized by various industries for decades, the AI boom is in full swing. Generative AI tools like ChatGPT and Google’s Bard have ushered in a new era of interest and innovation. It’s safe to say ChatGPT is responsible for piquing the general public’s interest. So much so that just two months after launching to the public, ChatGPT surpassed 100 million users. For comparison, it took Instagram two and a half years and Netflix 18 years to hit that milestone, respectively.
What many don’t realize is that while these generative tools steal today’s headlines, AI has been hard at work behind the scenes powering smart video surveillance solutions for years. For example, it has long been used in the public domain to detect abandoned packages, guns, altercations, and other indicators of crime. The benefit is the same across the board – the infraction has been identified before a human looks at the video, saving enormous amounts of time and money. Without AI, humans would have to examine hours of footage before seeing something worth investigating – if they catch it in the first place. So, AI is not only a huge time saver, but it does not get bored or tired like humans.
Regarding physical security, time and accuracy are the name of the game. If you can’t identify and respond to suspicious activity in real-time on video, is your security solution just for “after-the-fact” evidence? This is how record-and-store surveillance systems work – they capture footage to be reviewed later but don’t do anything to stop crimes in the act. And finding an infraction can be like searching for a needle in a haystack. Operators must look through hours, if not days, of footage to find the actual event. And when they do, watching a crime take place after the fact isn’t helpful. The thieves will be long gone with valuable assets and money in hand. To reduce the cost of crime, businesses need solutions that act in the moment to stop crime in its tracks.
Slashing Review & Response Times
Today, remote video monitoring (RVM) makes it possible for business owners to keep an eye on what’s happening in real time with real people reviewing and responding to suspicious events. By pairing video surveillance cameras with AI, we now have smart physical security solutions that alert suspicious events as they happen. As soon as the AI “sees” something, it notifies humans to examine what it found and to take further action. And because these systems are connected to the internet, users can pull up live camera feeds anywhere, anytime, on a phone or computer. Better yet, partnering with monitoring services ensures that trained professionals are always keeping watch – allowing business owners to focus on running their operations.
A common misconception is that AI is replacing people’s jobs. In fact, when it comes to physical security, AI is only helping humans do their jobs better. When AI detects an event like a car driving onto an empty construction site, it can alert operators to the event and they can determine if it’s worth investigating further, or just a false alarm. In Pro-Vigil’s case, AI has slashed the time it takes operators to respond to an incident from 20 seconds, in some cases, to less than a second. Because of this faster and more efficient process, human operators can focus on events that may pose an imminent threat and take actions to deter crime before it becomes a problem.
“Learning” What’s Suspicious
The best thing about AI in physical security is that it is always improving. Because high-definition cameras generate more data every year, they continue to help make AI smarter. Video captures everything from direction to speed, color, sound, and more.
These are all data points that can be used to train AI. As a result, models can be constantly updated as the system “learns” what to watch for and what to ignore. When it comes to physical security, a business isn’t concerned with cars driving by their perimeter fence after hours. However, if one of those cars makes a turn onto the property, then it’s something worth examining. Thanks to AI, that differentiation is possible.
Customizing AI By Vertical
Of course, it’s important to note that each business is unique. During monitoring hours, there are various activities that are expected and not expected. AI can make those distinctions. Because of this, there is no such thing as a “one size fits all” solution.
For example, customers may stop by an auto dealership after hours to check out inventory on the lot. But when it comes to critical infrastructure and utility sites, it’s rare that people or vehicles should be on site during monitoring hours. As a result, the smart surveillance system can determine what’s normal, and what might be a threat, for further examination.
Another example is a boat marina. The AI can be trained to know that movement from waves crashing against a dock is not a threat. Considering information specific to that industry and customizing the AI by vertical is vital in filtering out false positives to ensure better accuracy in detecting real threats.
“Machine Learning,” a subset of AI, can collect data points from cameras on various industry sites, enabling specific processes and AI models to be built to better serve a business’s needs. And once an RVM solution with AI video is up and running on a site, the model can even be fine-tuned for that specific location (or set of locations) for greater customization. The important thing is that AI is always learning and becoming more accurate.
Physical Security Powered By AI
For some, AI is just another buzzword. However, when it comes to physical security, the power of this technology can’t be overlooked. AI is changing the way businesses are secured around the world. By reviewing video faster than ever and reducing the number of false alarms, this technology is arming business owners and monitoring operations with the ability to detect, deter, and respond to crime more efficiently than ever. In a world where seconds make a difference, that’s quite a powerful tool.
Satish Raj is the Chief Technology Officer at Pro-Vigil, a provider of remote video monitoring (RVM) solutions powered by artificial intelligence. Pro-Vigil's leadership in RVM is enabled through proprietary technology developed by Satish's 75-strong engineering team whose expertise ranges from computer vision and at-scale video processing to IoT and cloud computing. Prior to Pro-Vigil Satish spent over 25 years building and leading technology teams delivering innovative software and hardware products.
