9.10.24 – SIW –Jason Lloyd
AI-powered technologies are providing administrators and police with new tools to mitigate acts of violence in our nation’s schools
This year marked the 25th anniversary of the Columbine High School shooting in the Denver suburb of Littleton, Colo., which claimed the lives of 13 people and left more than 20 others injured. Although there had been instances of mass shootings and other acts of violence on both K-12 and higher education campuses before Columbine, the massacre marked what many people felt was a watershed moment for school security. However, real substantive conversations on what was needed to address security gaps in our nation’s schools adequately were quickly bogged down in the political hot potatoes of gun control and arming teachers rather than placing the focus on risk mitigation.
Unfortunately, in the years since Columbine, there has been a litany of other school and community names that have been seared into the memories of everyday Americans. Sandy Hook, Virginia Tech, Uvalde, and Parkland: The list seems to grow with new names regularly added. And each time, this same process of finger-pointing and trying to assign blame plays itself out repeatedly while little is done to address real security threats.
The Search for True Innovation
Ironically enough, this uptick in school violence that started with Columbine has also coincided with foundational changes to security technology and the industry over the same time. For example, network video was in its infancy stages at that time, but today, deploying IP technology is standard for much of the market. There had also been little innovation regarding weapons detection, as one of the most common solutions for schools was and still is metal detection – be it of the walk-through or wand variety or some combination of the two. Today, however, spurred by 9/11 and other terror attacks around the world, there are a multitude of weapon detection modalities that can be deployed to prevent weapons from making their way onto campuses.
Additionally, during this time span, the first wave of video analytics rose to prominence. While they may have overpromised and under-delivered, this first generation of analytics laid the groundwork for this brave new world of video intelligence we find ourselves in today. Advancements in machine learning and neural network training have resulted in the development of what many people call artificial intelligence or AI-powered solutions today.
In fact, between the analog-to-IP migration, advances in mobile devices, the rise of cloud services, new weapon screening modalities, and AI, the industry has experienced a proverbial sea change in technology that is still underutilized in much of the education market. With that in mind, the real question becomes how school leaders can begin to take advantage of some of these advanced solutions while still navigating some of the challenges that have plagued schools historically.
The Elephant in the Room
The biggest barrier to increased adoption of security technologies in schools nationwide is budget resources or the lack thereof. Despite scores of lessons learned from the tragedies and the proliferation of video surveillance and other technologies throughout the K-12 and higher education markets, schools are still woefully lacking when allocating funds for security technology products.
Additionally, although school shootings have taken part in virtually every corner of the nation at schools of all shapes and sizes, the “it won’t happen here” sentiment is still a prevalent feeling among many school administrators. This is perhaps the most insidious of all the education market’s challenges. It helps instill a false sense of security and the belief that what they do is “good enough.”
The fact is that since the onset of the COVID-19 pandemic in 2020, where few if any, in-person classes took place and there was a subsequent dip in school violence, shootings at our nation’s schools are once again on the rise. According to CNN, school shootings spiked from 22 in 2020 to 73 in 2021 with another approximately 160 shootings occurring in 2022 and 2023 combined. As of June 24 there have been 34 school shootings reported throughout the nation so this year alone.
An Integrated Approach to Risk Mitigation
Of course, even for those districts with funds to spare, just throwing money at the problem won’t make it go away. School leaders must sit down with their security and IT personnel to develop a technology deployment strategy that makes sense from a budget perspective and the often-overlooked aspect of ease of use. One of the problems that many school districts across the country are dealing with is using a hodgepodge of different systems and devices purchased as needs arose or changed through the years.
Trying to manage these disparate technologies can be a nightmare for school administrators as many are proprietary or lack the adoption of different standards to make interoperability possible. Therefore, it is paramount that any security technology strategy being evaluated by a school system make openness and interoperability a key tenet of any system the district deploys en masse, be it video surveillance, access control, weapons detection, intrusion detection, or otherwise. Having a system that integrates each of these systems into a common operating picture is the Holy Grail of security technology deployments.
Trying to manage these disparate technologies can be a nightmare for school administrators as many are proprietary or lack the adoption of different standards to make interoperability possible.
With that being the case, it is also important that you be able to separate fact from fiction when evaluating the claims made by many vendors today about their offerings. Many people throw the term “AI” around when talking about their solutions, but how do you know they haven’t just sourced some analytics from a third party, slapped their name on it, and called it “revolutionary?”
Here are a few of the solutions you should investigate and the questions you should be asking about their capabilities and effectiveness:
Weapons Detection
Traditionally, screening students and staff in a typical school setting has presented several inherent challenges. First and foremost, adequate throughput is needed to avoid bottlenecks at entrances when children, teachers, and other staff members are arriving for the day. Although many districts have implemented clear bag policies to make the screening process quicker, trying to efficiently scan hundreds or even thousands of persons in a narrow window of time still presents significant issues. There are also still issues related to false positives, the time required to carry out secondary searches, and the expense of maintaining the equipment, which can be significant.
One solution many people are turning to today to address the shortcomings of metal detection is implementing newer concealed weapons detection technology, which allows people to move much more freely and naturally through them while maintaining high accuracy. Available from several different vendors, these solutions come in a couple different flavors – walkthrough portal and visual detection – and promise end users the ability to screen larger numbers of people quickly and more effectively.
Each of these solutions also touts its ability to leverage AI and deep learning technology to more quickly analyze and detect the presence of weapons ranging from small pocketknives to handguns, long rifles, and more. These solutions can also be seamlessly combined with existing video surveillance and access control management systems to create an integrated operating picture for the end user. However, it’s important to note that these systems should be thoroughly evaluated regardless of the vendor to determine their effectiveness and compatibility with existing security and network infrastructure.
Video Intelligence
As exciting as these recent developments in weapons detection are, they almost pale compared to the capabilities and intelligence that can be provided today via video. The advancement of neural network training and the ability to put thousands or tens of thousands of hours of training into an algorithm to refine its accuracy means schools can take advantage of a whole host of capabilities to bolster security and safety. Beyond weapons detection, schools can use various other analytics to become more proactive in their security posture.
Chief among these is fighting detection, which can provide school administrators, resource officers, and others with alerts about potential incidents and de-escalate them before they turn more violent. Of course, AI-powered modules can help schools keep track of a wide range of other potential threats, including fallen person detection, occupancy monitoring and crowd detection, objects left behind/removed, intrusion detection, smoke/vape detection, and much more. While each of these capabilities can be used to great effect on their own, when combined, they can create a comprehensive video intelligence solution that creates a force multiplier.
With staffing shortages remaining an issue for many schools even as other aftereffects from the pandemic become less commonplace, a system that can provide around-the-clock monitoring for these types of threats is critical for today’s campuses. Given the type of “high-trust” that these AI-powered modules provide, meaning that the alerts they generate can be trusted as if they were coming from a human operator, schools have never had this breadth of solutions available to help prevent violence and other incidents.
It is important to note, however, that not all so-called “AI” solutions are created equal, so it is crucial that you evaluate all of their claims to ensure that they can be delivered. Not all analytics providers are created the same. While some companies have rich pedigrees in analytics development and have robust portfolios of intellectual property, many others are simply white-labeling generic analytics pioneered by someone else.
Prohibited Person Detection/Faces as Credentials
Another challenge that has historically posed a problem for schools is being able to accurately identify those persons who are not allowed to be on campus – be they known sexual offenders or parents or guardians who have been forbidden contact with a child. However, the advent of facial recognition solutions means that schools can now tie their video network into their state’s known offender database to be alerted to the presence of these “prohibited persons” to be able to take immediate action. Implementing license plate recognition (LPR) is another effective way to keep banned individuals off campuses as security and/or police can be notified when the vehicle of a known offender is spotted on or near the campus.
For those looking to bolster their existing access control systems, solutions allow you to take an existing platform and convert it to a facial-based access system. Rather than using proximity cards or mobile credentials, authorized personnel can scan their faces to open doors around campus. What’s more, some of these solutions can take existing photos in the access control or visitor management database and apply them for access rather than having to re-enroll the entire badged population.
Regardless of the security technologies a school decides to deploy, they mean next to nothing if there isn’t a clear plan to leverage it fully and people are committed to ensuring the deployment is successful. The ability to detect all incidents is available to school leaders today, but it means little if policies are not adhered to or alerts are not followed up on promptly. So, while the dynamics and the technologies powering today’s school security programs may have changed, the principles that make them successful have not and will continue to be key as schools continue to evaluate technologies and attempt to stay on the cutting edge of security.
About the Author
Jason Lloyd | regional business manager for the eastern U.S. at ISS (Intelligent Security Systems)
Jason Lloyd is the regional business manager for the eastern U.S. at ISS (Intelligent Security Systems), a leading global video intelligence and data awareness solutions provider. Before ISS, Jason served as a regional sales manager for Vicon and headed up a low-voltage division at ADT for over a decade. He can be reached at jason.lloyd@issivs.com.