We’re hitting the phase with artificial intelligence (AI) where people start having realistic conversations about its benefits and risks. Up to now, AI has largely been spoken about as the solution for the world's unsolvable problems (curing diseases, enabling humanity to expand to Mars, ending poverty, and so on). Or, conversely, as the force that will bring about the end of civilization, starting with replacing people in the workforce and causing widespread unemployment.
Finally, the hype cycle is ending, and businesses are taking the tried and tested approach to applying groundbreaking technology: beginning with the end in mind. And for businesses, the end goal should always be to provide ever-increasing value to their key constituencies.
In the physical security space, AI promises to enable orders of magnitude higher value in terms of accuracy, scalability, and cost-effectiveness in protecting assets and preventing crime. This potential is a reason for optimism, as it can significantly enhance the industry's capabilities. However, the key for physical security customers is to become beneficiaries of AI and not frustrated victims. This article will discuss some considerations customers should consider when developing the optimal physical defense.
A Tool Like Any Other
It's important to remember that AI is not a standalone solution. Instead, it is a powerful tool or assistant that empowers humans to do previously impossible things and do them more efficiently. In the case of physical security, it makes live monitored surveillance commercially feasible while simultaneously speeding up incident response and reducing human error. This emphasis on AI as a tool should make you feel more in control of your security strategies, knowing you have a powerful assistant.
AI's ability to perform these tasks is akin to how human surveillance personnel gain expertise – through rigorous training in a specific environment. AI acquires this contextual understanding through 'machine learning,' where it is trained to recognize good and bad decisions. Context is crucial. For instance, AI can learn to accurately interpret an individual’s intentions by analyzing their behavior, the environment, the location's history, and other factors. All these elements contribute to the context in which AI operates.
AI Identifies Criminal Intent
Once AI is attuned to context, it becomes a powerful force multiplier in identifying criminal behavior. It can rapidly distinguish suspicious behavior from innocent behavior. It can “look” at an individual and determine if that person is behaving the way thousands of other criminals have behaved, and then determine the most effective way to avert a crime.
However, it is essential to set up the surveillance process so that the ultimate accountability for preventing crime always lies with the human surveillance experts. This ensures that AI is used responsibly and that its immense potential translates to substantial value for the customer.
AI Does Not Work Well “Out of the Box”
As mentioned earlier, AI needs to be “taught” before it can become useful—just as humans need the “experience” of repetition” before they can master something. This is known as AI training. AI can identify similar behaviors in actual operations by repeatedly processing incidents during training.
This is one area where some organizations have a negative experience with AI – they expect it to work “out of the box” and discount the importance of training. With remote video monitoring services (RVMS), this amounts to not engaging with the right service providers. The right service providers could train AI through substantial crime prevention experience. They take direct responsibility for preventing crime and view AI as an enabler, not the solution.
Again, clearly defining the end goal(s) and priorities is important. For instance, the primary purpose of AI in RVMS is typically security. That is not to say the data being ingested cannot be used for other purposes (for example, site traffic analysis), but it is important to use AI to solve a defined problem first and then branch out from there. This approach ensures that AI helps solve the intended problem first and, moving forward, delivers added value.
The Future of AI in Physical Security
AI is still relatively new in physical security but is already helping solve immediate problems – stopping crime in real-time and making RVMS feasible for various scenarios.
Moving ahead, AI will evolve from a real-time crime-stopping mechanism to a tool that enables us to predict crime with increasing accuracy and specificity, as there are virtually no limits to AI’s capability to learn and extrapolate with experience. This can transform physical security by ensuring that scarce resources (particularly on-site guards, law enforcement personnel and remote video surveillance experts) are focused on the right situations. Likewise, as we mentioned, AI can be expanded from security to providing value in other areas – wouldn’t it be good to use predictive AI to, for example, assist you in making accurate staffing decisions by watching foot traffic?
AI is one of the most impactful technological breakthroughs of our time. It can enable us to realize previously unimaginable value. For businesses, it is important to keep things in perspective – AI is a tool. It is not a cure-all out of the box. As with any tool, it works well when used responsibly by the right experts whose focus is providing value to customers over anything else.
