*Originally published on SecuritySales.com
Illusory marketing hyperbole no longer defines AI, machine learning, business intelligence and the like in the physical security realm. Get a brass tacks perspective where broad adoption is happing today and into the future.
Artificial intelligence (AI) is a growing force in the physical security industry. That declarative statement may well have you grinning, wryly, especially if you’ve been exposed to the near carnival-barking level of embellishment at trade show exhibitions the past few years. But in reality, the statement could not be truer.
SSI spoke with a number of vendors that specialize in this developing technology arena to uncover the ways organizations are approaching and implementing emerging technologies in order to support security and business initiatives.
Read on as these thought leaders offer a state of the AI landscape, including the latest developments in security and safety applications, early adopter markets, integrator opportunities, longer-term projections and more.
Newfound Value Proposition
A tell-tale indicator of the evolving security ecosystem is how technologies such as AI, predictive analytics and machine learning are fundamentally changing the core value proposition of physical and electronic security.
Traditionally, access control, video surveillance and intrusion systems have been highly reactive, says Blaine Frederick, vice president, product & engineering, for Alcatraz AI, an autonomous access control platform provider based in Redwood City, Calif.
He uses the example of a security officer who might be monitoring a camera or an alarm panel and responds when there is an issue. This means the officer potentially needs to be able to act in response to multiple issues simultaneously.
The real-time predicament is intensified when the officer is dealing with multiple buildings and in multiple time zones, as would be the case in a global security operations center (GSOC) of a Fortune 100 company.
The customary resolution would be to add more officers to scale up. However, today there is another option: autonomous experiences that allow companies to rethink staffing plans without sacrificing security.
“Machine learning technology has allowed manufacturers to create products that can not only aggregate the data into easily digestible dashboards, but can also bring insights into trends like when are most of the alarms coming in, what locations are the most problematic and can then automatically redirect resources in real-time to deal with issues,” Frederick explains.
The data then becomes valuable to other stakeholders within the company beyond just the security team. This fundamental change is transforming the security group from a cost center to a contributor potentially as a profit center.
“Alcatraz has taken this approach one step further by applying these predictive techniques to face authentication, tailgating and compliance [mask] enforcement. The offering is unique in that we deliver these analytics at the door using edge computing,” Frederick says. “This delivers an access control experience that rivals having a live security officer at each door checking IDs and ensuring people are not sneaking in.”
During the past two decades, the amount of data needed by a business to operate successfully is growing at an exponential rate, says Aluisio Figueiredo, CEO of Intelligent Security Systems (ISS), a global provider of video management and image analytic software headquartered in Woodbridge, N.J.
Moreover, there are fewer and fewer highly qualified personnel capable of performing qualitative analysis of these data sets. And that is where AI comes in. Figueiredo says the high demand for AI-enabled products is determined by two factors:
- The development of computers that are able to collect, correctly classify, analyze data comprehensively, draw logical conclusions revealing previously unexplained patterns, and as a result, forecast and make well-founded decisions many times faster and better than a human.
- The business desire to have steadily increasing profitability and to be competitive with limited HR and time resources.
“Nowadays the computer allows almost instantaneously performing a lot of tasks that would take a person hours and even days,” he says. “This is not only about classical security tasks, but also about a wide range of challenges related to human behavior in the workplace, in the office, in transit — wherever noncompliance with conduct standards and regulations may lead to economic loss or threat to health and life.”
After years of overhyping and deceptive marketing, AI is finally beginning to be more widely adopted and performing according to its promise as the computing side has matured, says Srinath Kalluri, CEO of Oyla, a computer vision-based AI company located in San Carlos, Calif.
AI-driven video analytics, for example, are finally becoming robust enough to be widely deployed in the field, driven by an increasing array of low-cost sensors, edge-computing plat-forms and Cloud-based architectures.
“The fusion of technologies, like video and AI, is most promising for the security industry from my point of view. This ‘fusion’ is the combination of advanced sensing technologies unitized physically and in software to create a solution that ultimately provides better awareness in a single device or series of devices,” he says.
Additionally, Kalluri continues, the proliferation of Cloud-based technology and edge-based capabilities that are being used to remotely access and monitor video are evolving quickly as enabled and marketed by Intel, NVIDIA, Qualcomm and others.
COVID Fuels Adoption
Fortuitous may not begin to explain the role the coronavirus pandemic has played in advancing AI-enabled solutions in terms of creating awareness and acceptance for these technologies in the physical security realm. Simply, COVID-19 mitigation and business recovery efforts have greatly accelerated adoption of a variety of AI-enabled solutions.
“Work-from-home mandates meant many security personnel couldn’t physically go to their place of business, so they had to implement and rely on AI-powered video surveillance solutions and other AI-enabled technologies to take on the responsibilities they could no longer do,” says Jeremy White, founder of San Antonio-based Pro-Vigil, a provider of mobile surveillance units, live video monitoring, event-based surveillance and IP health monitoring. “These technologies also served as a force multiplier for companies facing labor shortages due to employee illness, quarantine requirements and occupancy limitations.”
The pandemic also introduced new use cases for AI. Many organizations are now leveraging AI technology to monitor compliance with Centers for Disease Control and Prevention (CDC) COVID-19 prevention guidelines, White says. “Pro-Vigil is leveraging AI in just this way in its Social Distancing AI Suite, which helps customers measure compliance with social distancing requirements, occupancy limits and facemask usage.”
The pandemic has caused a shift toward touchless technologies and specifically touchless access control solutions. Although touchless access control technologies have existed for a number of years, the core technology powering them has had some inherent flaws, explains Frederick.
For example, traditional biometric identification has used statistical probability calculations to determine if a subject interacting with a device is the same as a particular record in its database.
“This approach can be accurate; however, it does not allow for changes in the reference record [enrollment]. In face authentication, things like gaining weight or losing weight or growing a beard or shaving a beard have historically caused problems with authentication making the systems difficult to use,” he says. “By introducing deep neural networks the reference records can be updated over time, which significantly increases the tolerance for changing appearances over time.”
Creating products that adjust to real-world scenarios like changing appearances allows touchless access control technologies to become more widely accepted as the user experience — reliability and ease of use — is greatly improved, Frederick adds. While the pandemic has created more security awareness around the world, the United States in particular has been responsive to new service offerings, says Ryan Schonfeld, founder and CEO of El Segundo, Calif.-based RAS Watch.
As a provider of high-tech GSOC services, the company has seen changes being implemented across all different kinds of organizations, especially in healthcare and education sectors. “We’ve seen a shift towards more SaaS solutions that work remotely, such as remote access control, and secure infrastructures across multiple sites, such as doctors meeting with their patients through Zoom-like infrastructure,” Schonfeld says. “We will look back and notice that the pandemic pushed AI into more mainstream acceptance, as organizations had to quickly find robust, innovative and effective ways to secure their employees and assets.”
Kalluri says one simply cannot discount the impact that COVID-19 has had on the market — especially for security systems integrators. The pandemic has disrupted so much of our lives, but for security integrators, the disruption has been around the normal cycle of deployments and the uncertainty influencing the buyer’s cycle.
Yet, at the same time, the market for emerging technologies continued to grow over the past year, driven forward by an increasing focus on technology development — primarily around driving more intelligence-based video surveillance to customers. Advances in AI have added more value to analytics solutions, which have traditionally been plagued with an inability to deliver true, real-time intelligence.
“Trading simple rules-based approaches for more deep learning-enabled functionality realizes the promise of this technology and better meets end-user expectations as they increasingly need to manage by exception,” Kalluri says.
By tradition, Kalluri adds, as social anxiety rises the security industry performs better as companies prepare to protect assets and people from any unrest. He projects a robust market in 2021 as society emerges from the pandemic and life returns to some semblance of normal.
“Additionally, companies thinking more strategically based on lessons learned will realize the benefits of remote management, enhanced intelligence capabilities and new ways of doing business,” he says, “and all of these things are supported by the advances in AI and security intelligence gathering.”
AI on the Horizon
When Figueiredo takes a long view of the physical security ecosystem, he sees two industries in particular where AI technologies could make major impacts. The first industry is manufacturing facilities. He explains that labor safety and compliance with safety requirements are a priority since the outcome of people’s work is what determines the productivity and quality of the final product.
The AI video analysis systems help ensure a high level of personnel control and minimize the number of hazardous situations. Contemporary neural network-based video analysis systems ensure compliance with labor regulations and health standards, monitor people’s physical wellbeing and control the wearing of personal protective equipment, such as safety helmets.
These systems, he explains, in addition to identifying patterns of dangerous and irregular behavior also inform supervisory officers of violations, collect statistics and provide efficient and user-friendly data management.
“Nevertheless, the use of AI technologies improves product quality, increases the output and reduces the percentage of rejects, minimizes the possibility of delays due to production incidents, and optimizes the operating costs, that is, enhances the reputation and improves the enterprise’s economic efficiency,” he says.
The second industry cited by Figueiredo is intelligent transport systems. AI-powered solutions provide auto-traffic control and rapid detection of road accidents and incidents that could lead to emergencies. Moreover, using AI can enable operational and long-term forecasts; for example, about traffic congestion, the probability of an emergency on a road section, among other projections.
White at Pro-Vigil explains AI systems will become significantly “smarter” and more accurate this year because they will accumulate a critical mass of data for machine-learning training, much like how young humans become smarter with each year of experience.
Because of this, video surveillance systems that leverage AI functionality will move beyond mere object recognition (“there’s a person on my property”) to predictive deterrence (i.e., stopping a crime before it happens) using a predefined set of criteria.
“For example, if someone emerges from the shadows carrying a bolt-cutter near a chained fence gate, it could notify a business owner or even turn on lights and make announcements to deter the person from attempting a break-in,” White says.
Alcatraz AI’s Frederick says he expects the industry will continue to see a convergence of physical and cybersecurity moving forward. The number of security vulnerabilities and breach attempts will continue to increase. Systems that utilize AI can begin to learn/predict what potential threats will be able to prevent/mitigate them before attackers can implement them, he explains.
“With an increase in connected devices and applications, enterprises are becoming more vulnerable as they are connected to a plethora of independent endpoints,” Frederick says. “Protecting physical security end-points on the edge via AI provides a compelling proposition with its proactive threat mitigation capabilities, which are needed for constant supervision and adaptation to the multifaceted security vulnerabilities faced by the modern digitized enterprise.”
Schonfeld at RAS Watch also projects a promising future for increased adoption of AI-enabled technologies, with an expanding base of end users even far removed from super-scaled, enterprise-level systems. The pandemic may have helped usher in more Cloud-based security solutions and advanced intelligence, but what the industry has experienced so far is just the beginning of broad adoption of these practices, he says.
“SaaS and AI are well on their way to being fully utilized in mainstream sectors,” he says. “While Cloud-based solutions, remote access control and AI-driven video monitoring are already widely used in places like airports and government facilities — because of the scalability and ease of transition — we will begin to see the shift of these solutions taking root in your local high school, college campus, healthcare
clinic and more.”