The emergence of sophisticated artificial intelligence technology promises to be disruptive to many industries, including security. That much was clear with the number of exhibitors at ISC West boasting some kind of latest-and-greatest AI tool or innovation.
Some of the most innovative minds in the security business shared their insights about AI during a SecurAmericas Forum at the show, hosted by video intelligence and data awareness provider Intelligent Security Systems.
The event brought together thought leaders from across the security industry to discuss the state of AI in video surveillance applications and how the adoption of the technology will continue to evolve throughout the market.
ISS Chairman Richard Burns kicked off the discussion by calling attention to ChatGPT and how it is reshaping how organizations think about and leverage AI in day-to-day use cases.
Security leaders at the roundtable acknowledged that security integrators will have to beef up their knowledge of the technology to evolve with the industry. If they don’t, these firms could see their market share shrink due to potential competition from companies capable of using AI tools to help customers solve problems.
Eli Gorovici, a founder of NICE Systems and DVTEL and former general manager of Johnson Controls Security Products, sees applications for AI in healthcare, the financial world, cybersecurity and more – especially with risk management.
“I think is really going to revolutionize our world, our lives, our work. Most of our end users don’t care about the video or the access control. They’re trying to mitigate risk. And if you really take AI into these places, this is where a huge memory system can do amazing things,” Gorovici says.
For example, an integrator may be working on a system installation and the architect or engineer uses AutoCAD software. The design could take a few weeks or more, “but you can throw it into an AI system and in a few hours, you finish the design,” Gorovici says. “It’s not just video analytics. We have the design now. It’s not just placing cameras and access control, but also looking at the environment and the city, the location and the crime rate and truly having risk management to the site.”Challenges of Deployment
Bill Bozeman, the former president and CEO of The PSA Network, says integrators made an excellent transition adjusting to the needs of enterprise-level access control. But he sees a more challenging road ahead for traditional integrators deploying AI technology.
A few years ago, Bozeman says he attempted to introduce “incredible” opportunities in the cybersecurity market to a group of 200 integrators that he worked with, but it was a “dismal failure.”
“I don’t believe that would be the case here,” Bozeman said of AI deployment. “Are (integrators) going to have a choice? What’s going to happen if they choose not to receive an education and choose to stand on the sidelines? There will be a spot for the smaller integrator. It won’t be a very impressive spot. But I do believe the smart security integrators will step up to the plate and deploy.”
He adds that some integrators struggle with deploying new technology because they’re busy and comfortable.
“It takes a larger organization to get their hands around the challenge of deploying, and to have the finances do the training,” Bozeman said. “If you’re already fat and happy and a small company, it’s tough. Either we’ll need to have AI products that are very, very easily deployed, or have a network of high-end integrators who get the training, and they use the smaller guys for more basic things.”
Monetizing Data
Matt Kushner, the executive advisor to Stanley Security (now Securitas Technology) and former President of Stanley Security Worldwide, said there are hundreds of billions of data points being consumed through the systems integrators deploy, but it’s not all being utilized.
“But we’re an exception-based industry. We monetize an alarm event. So, 3% of what happens is what we monetize. The other 97% of data sits dark but would give you unbelievable insights to what's happening in a building,” Kushner says. “It’s highly contextual data that tells you what a building looks like in a normal condition, and it can identify an abnormal event.
“We’re an industry that reacts to crime, but we have data sitting there that would allow us to be a proactive industry to forewarn that we’re vulnerable,” he continues. “Are we the industry to do it? The question is, ‘Are we funded to do it?’ Yes, we're busy, we have customers, and we have shareholders. We pull gross margins down to 34%. That’s a good deal. The competitors that we’re going up against have gross margins that sit at 70%. We are at the bottom of the value pyramid.
“The folks that will come into our space don’t roll a truck and don't have this low-voltage technician with carpentry skills. They don’t want it. They want pure data solutions, and they have their crosshairs on our industry. Trust me.”
‘A Rising Tide’
Pierre Trapanese, owner and CEO of Northland Controls, notes integrators are closest to the customer and potential competitors don’t speak integrative languages very well.
“They don’t know how to deal with things in the field, so they're highly dependent on the integrators,” he says. “If we have these intelligent devices out in the world trying to do stuff, and it must be calibrated, that’s a hands-on affair.
“The integrators must be a part of this process. There will be integrators who will lag and there will be a place for them as the market moves because it's a rising tide. I don’t think it’s a tidal wave.”
One of the challenges integrators will face is figuring out the “multiplicity of things you can do once you put the camera in,” says John Mack, an executive vice president of the Investment Banking Group at Imperial Capital.
“For example, retail analytics. If you’re deploying video in a retail environment, the retailer expects you to do all kinds of things with intelligence around what's happening with traffic patterns and dwell times,” Mack adds. “Well, security guys aren't necessarily experts in retail analytics.”
Another example he provides is Walmart, which does a lot of work with automated inventory management. “There’s a ton of AI involved there. Is the security integrator the expert in deploying AI to work with the video with an inventory management system?” Mack says. “It will be a really interesting challenge to figure out how the security integrator can be in touch with these other constituents who will be part of using the camera for AI applications.”
Solving Problems
The pressure on integrators now may be different than in previous technology shifts, such as when IP cameras came to the industry, admits Matt Powell, managing director for North America at ISS.
And there are additional challenges for the security industry in defining the language and differences between AI, analytics and machine learning or neural networking.
“Five years from now, we will have integrators that have no idea what a Mercury board is because they are just going to solve problems utilizing the infrastructure that everyone in this room installed,” Powell explains. “They won’t know what a Mercury board is. They won’t know what an IP camera is. They will literally look at what’s installed out there and say, ‘We don’t care, give us an API connection and we’re going to solve problems for you, while we’re talking about product offerings.’”
Steve Surfaro, recognized industry technologist and chairperson of the Security Industry Association’s (SIA) Public Safety Working Group, contends AI is for people that want to make decisions faster and do predictive modeling.
“Those are many of the efficiencies that we can accommodate, but we won’t understand that unless we understand some of the hardware challenges that are occurring,” he says. “Right now, most of the solution providers in the industry are looking to edge AI, which is absolutely ridiculous. You should be focusing on offloading the CPU cycles from your VMS platform and using an AI processor.”
Workforce Re-Training
Sameer Sharma, general manager of IoT and Smart Cities for Intel, said the three biggest buckets his company sees AI being applied to is public safety, transportation and sustainability implementation, which is a good position for the security industry to be in.
But he warns about the speed at which AI development is occurring and how quickly it’s changing the economy.
“Three weeks ago I was chatting with a small group of people who were on the leading edge of AI development, including some of the early seminal work, and one of the people that I really respect -- I mean they cannot predict anything -- but what they’re forecasting is in about 3 years a typical programmer's job will be done by an AI bot,” Sharma concludes.
“That AI bot is going to cost you about ten dollars a month. Think about the upheaval that's going to cause in terms of our workforce retraining. We must think about workforce re-education.”
John Dobberstein is the managing editor of SecurityInfoWatch.com and oversees all content creation for the website. Dobberstein continues a 34-year decorated journalism career that has included stops at a variety of newspapers and B2B magazines.
