What Can AI Do for You?

The security industry has no shortage of buzz terms, and one that is consistently in the discussion is artificial intelligence, or AI. I’ve heard about AI since I was a kid – as early as The Terminator’s Skynet and HAL from 2001: A Space Odyssey, the concept of AI has been presented to us as a science fiction concept that will inevitably turn on us and destroy humanity. While present-day incarnations of AI are less devious, the aura of science-fiction still looms, as if they are some cosmic tool we cannot understand but desperately want to deploy in our security systems and operations. The present reality is less fantastic; however, the application of AI can be beneficial. This article will define AI, provide additional context to the different types and discuss its current application in the security industry.   

What Is Artificial Intelligence?

The dictionary definition of AI is the theory and development of computer systems able to perform tasks that normally require human intelligence. It provides a vast spectrum of activities that would fall into this category. At a basic level, AI automates tasks that would require human intervention or decision making. This could be as simple as your video systems providing notification when an object appears in the field of view or more complicated, like scanning through all your video to categorize the footage with red cars. 

As we move up levels of complexity, we graduate into the realm of machine learning, or ML. It is a subcategory of AI, where the systems are trained to adapt without following direct human instructions. The systems leverage algorithms and statistical models to analyze and draw inferences from patterns in data. The programmer sets the general parameters or rules; however, the system is free to make connections with the data and draw conclusions. We see this applied in facial recognition and authentication – the system can identify an object or a person and, through scanning the image multiple times, learn more about the person. How do they look in glasses? When wearing a mask? How does the side of their face look versus a front view? Based on this information, the system can make decisions regarding allowing access or notifying someone of a VIP’s entry into the facility. With ML, we can provide structured data sets and program the systems to make decisions based on the information it collects. There is yet another level of AI, referred to as deep learning– this is where we get much closer to the self-aware drones and the need for Neo to save us from the Matrix.

Deep learning is a further subset of AI and ML, where the system is much less reliant on human intervention to structure information and provide training. The system can observe patterns in unstructured data and categorize the data – it can monitor and make sense of what is being viewed. Deep learning is most valuable when ingesting large amounts of data to provide greater accuracy levels – the more information it has, the better it can learn. Some examples of uses cases are digital fraud detection and autonomous vehicles. 

AI in the Security Industry 

Video management has been the most prevalent adopter of AI in the security industry due to its great applicability with image recognition and categorization. One of my favorite use cases is image searches for investigations – gone are the days of pouring over video footage to see when the black sedan pulled out of the parking lot or when a particular person left the building. With this feature, operators can ask the system to search for similar images and sort through a half-dozen instead of hundreds, saving significant time and resources. As programs mature, we see more systems connected to create a more data-rich environment where end users can take more substantial advantage of AI. This is where exciting opportunities exist and the potential for significant innovations in our industry. 

What Is Stopping Us?

In my experience, the greatest challenge for the deployment of AI for a security organization is their level of ‘AI-readiness.’ In the white paper The AI Readiness Model, Intel Corporation provides an excellent framework for the considerations for organizations inspiring to deploy AI. In the framework, they discuss the foundational programs readiness factors such as:

• Does your organization already have any AI tools in place?
• What data will you use?
• Where is the data?
• Who owns it, and how will you gain access?
• What resources are needed to deploy, then maintain?

The punchline here is that it is not as simple as buying a single security product that “uses AI,” plugging it into your network and voila, you got yourself some artificial intelligence. Like any other program deployment, it requires a thoughtful and well-planned approach. 

With that said, I am bullish on the future of AI in our industry. In particular, I would encourage young professionals to invest time in understanding how it works, along with the application to security technology. Everything from incident management to customer service can be optimized to some extent by the application of AI. Your ability to understand how it works, the applicable use cases and the value to security organizations will become a competitive differentiator.

The views and opinions expressed in guest posts and/or profiles are those of the authors or sources and do not necessarily reflect the official policy or position of the Security Industry Association.