Q&A: How AI is Impacting the Security Industry

The creation of generative artificial intelligence (genAI) is one of the most significant technology inflection points in history. AI developers continue to make great strides, releasing new algorithms to power natural language processing (NLP), computer vision, and further genAI capabilities.

“These advancements position AI technologies as one of today’s most transformative technologies, with the potential to reshape industries and society by automating complex tasks at scale, enhancing conceptualized decision-making, and enabling increasingly personalized capabilities,” according to a whitepaper series on AI by the World Economic Forum.

Ahead of GSX 2025, we reached out to the ASIS International Emerging Technology Community Steering Committee Chair Quang Trinh to learn more about how AI is impacting the security industry.

The transcript below has been lightly edited for clarity.

GSX Daily. What are some of the most important developments in AI in the past year that security practitioners should be aware of heading into GSX 2025?

Trinh. The most significant impact is on computer vision. The shift from pixel-based motion algorithms when we transitioned from analog to digital was a major leap towards more efficient use of the system, particularly with video motion detection (VMD). Now with AI, architectures in deep learning AI models and new ones—such as contrastive language image pre-training (CLIP) and vision language models (VLMs)—we are seeing a baseline shift to standard object-based detection and classification capabilities, with many vendors leading the way with people and vehicle classifications as the new baseline.

VLMs are a multi-model approach to language, images, and videos, and will usher in new capabilities that make configuring, optimizing, and using security systems much easier and more natural. These interactions will help us gain a better understanding of both human input and the translation of the real world into a digital format.

Another rapidly evolving architecture is agentic AI, where AI agents are created with a specific profile that can assist a person interacting with them. Enhancements in operations through Standard Operating Procedures (SOPs) have a significant upside in automating repetitive tasks with agentic AI. This also helps address training and workforce resource gaps—a person can get up to speed more quickly with an AI agent designed to understand the organization’s operations and processes.

GSX Daily. What’s a recent security use-case of AI that stands out to you as a potential game changer for the industry?

Trinh. A lot of emphasis is placed on security, but AI-based systems are accelerating the aspect of safety and providing a safer environment for internal and external stakeholders. Many components of a security system are hundreds, if not thousands, of sensors. These include IP cameras, thermals, temperature, humidity, light, acoustics, and access control checkpoints for readers and gates. These sensors can now provide information and understanding of risk and safety compliance.

Customers in manufacturing are discovering that computer vision advancements can detect and identify falls, spills, and other abnormal activities from both humans and machinery within a manufacturing plant, allowing them to address issues proactively rather than retroactively after an incident has occurred. The cost of business continuity justifies the use of AI technologies in parallel systems to provide more than just security, but also safety and operational efficiencies.

More than just interacting with the IT team, other business units within the customer organization will also have an influence on expanding the security system's functions. This could include teams in operations, logistics, marketing, sales, HR, and others that have a vested interest in the current security infrastructure and the expansion of its capabilities, as well as those of other systems.

GSX Daily. What are some of the risks related to implementing AI-enabled solutions that security practitioners should be mindful of?

Trinh. Trust but verify. There will be a lot of solutions being presented to them from vendors and companies trying to sell a solution. End customers and integrators should be familiar with the capabilities, but they should also vet and test the solutions before committing to them.

AI technology itself is not the solution, but rather an enabler that helps a solution work and execute more effectively. I would caution against marketing messaging that portrays AI itself as the silver bullet for the solution.

Another aspect will be regulation and compliance, which includes local, state, federal, and international compliance. A lot is happening in the regulatory space, and it is changing every day. I would encourage people to ask questions regarding regulation, starting with data privacy and how these solutions address data privacy before discussing the AI architecture itself.

GSX Daily. What areas of AI is the Emerging Technology Steering Committee paying particularly close attention to or researching?

Trinh. We are open to all sub-categories under AI, including, but not limited to, computer vision, language, acoustics, UAVs, and robotics. Many of the foundational technologies will impact video surveillance, access control, risk mitigation and assessment, and intrusion detection systems.

There are breakthroughs in geospatial data and digital twin applications that also positively impact our industry.

GSX Daily. The Emerging Technology Community is sponsoring 23 education sessions at GSX 2025—including 15 that touch on AI. What guidance would you give attendees on deciding which of these sessions to attend?

Trinh. The ones that we sponsor and endorse incorporate elements of emerging technology and their applications to various sectors of our industry. I would recommend sessions that provide practical uses of the tech in your environment, which we tried to highlight.

It is easy to talk about the negatives or even the future potential use of AI but applying practical environments and uses into the equation provides the audience with a better perspective on how it is impacting them today.

GSX Daily. Is there a particular session or speaker that you’re looking forward to most at GSX 2025?

Trinh. Some notable sessions on my list are:

Monday, 29 September 2025

• “What AI Really Looks Like in Security Operations”, 11:15 a.m. to 12:15 p.m.
• Panelists will discuss the ins and outs of AI in practice, explore integration of AI into security operations, and highlight its potential role in surveillance, detection, and predictive risk analysis.
• “AI Security Convergence 2025: Navigating the Integration of Physical, Cyber, and Commercial Security Intelligence”, 2:45 p.m. to 3:45 p.m.
• Michael Askew, owner and managing director at Monitor Computer Systems, will share how AI technologies have evolved since their early adoption to address current challenges and opportunities in 2025.
• “Agentic AI: The Inflection Point for Corporate Security”, 10 a.m. to 11:00 a.m.
• Chris Hurst, vice president of global intelligence at Crisis24, will explore agentic AI applications, including threat detection and mitigation for physical security, coordination and decision-making in crisis management, and personalized solutions for travel risk management.
• “AI Evolving: How Harnessing Agentic AI for Video Drives Smarter Security Solutions”, 1:30 p.m. to 2:30 p.m.
• Barry Norton, vice president for human-centered AI, Milestone Systems, will discuss the foundational distinctions between large language models and vision-language models (VLMs), the transformative role of VLMs in video security, and how agentic AI is poised to revolutionize security applications by creating smarter, more proactive solutions.

Tuesday, 30 September 2025

• “Enhancing Security Screening with AI: Transforming Threat Detection Across Industries”, 1:30 p.m. to 2:30 p.m.
• Amir Lidor, vice president of products at SeeTrue, will explore the transformative role of AI in enhancing security screening and redefining threat detection across industries.
• “Unmasking AI-Powered Fraud Tools on the Dark Web: Threats and Countermeasures”, 11:15 p.m. to 12:15 p.m.
• Peter Warmka, CPP, founder and CEO of Counterintelligence Institute, will share real examples of how AI  technologies are being used to conduct fraud at unprecedented scale.
• “Navigating the Drone Dilemma: Enhancing Security While Managing Risks”
• This panel session will examine the dual impact of drones, exploring benefits like enhanced monitoring and response times alongside the challenges they pose to physical security planning.

Wednesday, 1 October 2025

• “Revolutionizing Video Monitoring: AI Integration for Enhanced Security Operations”, 9:45 p.m. to 10:45 p.m.
• Panelists will explore how AI is transforming video monitoring while strengthening traditional security practices, including through live demos.

GSX Daily. When it comes to learning about AI, what are some of the resources you turn to most and would recommend to other security practitioners?

Trinh. There are a ton of resources available that are free and premium. I recommend looking at freecodecamp.org. It has a YouTube channel where it hosts the complete Harvard Computer Science Course CS50 online.

There are platforms like Cousera and others that provide training, as well as certifications for individuals. Industry associations like ASIS and SIA have emerging tech and AI groups that provide information and insights into AI’s impact and influence on the security industry. People can also leverage training from vendors because they might have materials and courses that are relevant.

Megan Gates is senior editor with Security Management, which publishes the GSX Daily. Connect with her at [email protected] or on LinkedIn.