AI Is Moving Faster Than Our Security Playbooks
At GSX 2026, one keynote challenged the audience to think beyond incremental change.
Peter H. Diamandis described an “exponential world ahead,” driven by advances in artificial intelligence, computing capacity, data, automation, and capital investment. Some of his forecasts were intentionally bold. But the most useful message for security professionals was practical: the pace of change is accelerating, and our security programs cannot remain designed for a slower world.
For those of us in physical security, this is not only a conversation about new cameras, analytics, or software. It is a conversation about how we assess risk, make decisions, support customers, govern technology, and preserve human accountability.
The Question Is Not Whether AI Will Affect Security
AI is already influencing how organizations collect information, detect patterns, prioritize events, produce reports, develop proposals, and respond to incidents.
The opportunity is significant. AI can help security teams process large amounts of information faster, spot trends they might otherwise miss, and reduce time spent on repetitive tasks. For integrators and security leaders, it can also improve design decisions, estimating, project documentation, service knowledge, customer communication, and operational follow-up.
However, speed alone is not value.
A faster process that produces inaccurate recommendations, uses poor information, or operates without accountability can create more risk instead of reducing it. The goal should not be to add AI to every process. The goal should be to identify where it can help people make better decisions and deliver better outcomes.
That is where professional judgment remains essential.
AI Should Strengthen Human Judgment
One of the keynote's strongest points was that AI agents are becoming more capable. An agent is more than a chatbot that answers questions. It can be assigned a goal, use approved tools, analyze information, and recommend or take defined actions within its permissions.
This can be useful in security. An agent may help review service-ticket history, identify recurring equipment issues, organize project information, compare design alternatives, or prepare the first draft of a risk summary.
But treat every agent as part of the security environment.
The keynote presented three simple governance principles that apply well beyond cybersecurity:
- Every agent needs a clear identity, a responsible owner, and an expiration or review date.
- Access should follow the principle of least privilege. Start with read-only access; any ability to change information or take action should be deliberate and controlled.
- Activity should be auditable. Organizations need to know what the agent accessed, what it recommended or changed, and who approved it.
These principles are familiar to physical-security professionals. We would not provide unrestricted access to an employee, contractor, or system without knowing who they are, what they can access, and how their activity is monitored. AI should be managed with the same discipline.
Physical Security Is Becoming More Connected
The keynote also highlighted the growth of connected sensors, autonomous systems, drones, robotics, and machine-generated information.
For physical security, this means that more of the environment will be observable. Cameras, access control, intrusion detection, environmental sensors, intercoms, vehicles, and other systems will produce more data.
This can create better awareness. It can also create noise.
Value won't come from more data alone. It will come from connecting the right information to the right decision at the right time.
A well-designed security program should help an organization answer practical questions:
- What happened?
- What does it mean?
- What requires attention now?
- Who needs to make the decision?
- How does this action reduce risk or support operations?
This is where system design, operational knowledge, and customer understanding matter. AI can help organize and interpret information, but it cannot replace a clear understanding of the organization’s mission, priorities, risks, and constraints.
Start With a Real Business Problem
The best AI initiatives will not begin with a technology demonstration. They will begin with a real operational need.
For example, a security integrator may use AI to help estimate labor more consistently, search historical project information, identify common service issues, or prepare a first draft of a customer-facing summary. A security leader may use it to organize incident trends, improve report quality, or identify recurring vulnerabilities across facilities.
In each case, the question should be the same: does this improve a meaningful outcome?
The keynote offered a useful framework: map the process, teach the system, keep humans responsible for decisions, and measure the result.
Before implementing an AI tool, I recommend asking:
- What specific problem are we trying to solve?
- What information will the tool need, and is that information appropriate to use?
- What decisions may the tool support, and which decisions must remain human?
- What access does the tool require?
- How will we measure whether it is improving quality, time, cost, risk reduction, or customer experience?
Clear answers to these questions are more valuable than a long list of AI tools.
Curiosity Is Now a Professional Responsibility
The keynote also emphasized mindset: purpose, curiosity, an understanding of exponential change, abundance, and the willingness to pursue ambitious ideas.
For me, curiosity is especially important.
We do not need to become experts in every AI model or every emerging technology. But we do need to keep learning. We need to ask better questions, test ideas responsibly, and understand how new capabilities may affect our customers, our operations, and the security profession.
A useful exercise is to examine a process we know well and ask: If I had to redesign this process today, using the tools now available, what would I do differently?
That question can lead to valuable improvements. It may reveal a repetitive task that can be automated, a decision that lacks information, a workflow that is too slow, or an opportunity to provide better service to a customer.
Technology Changes, Responsibility Remains
The future described at GSX may feel fast and, at times, uncertain. But the core responsibility of the security professional does not change.
We are still here to protect people, support operations, manage risk, and help organizations fulfill their mission.
AI can make us more effective. It can help us spot patterns sooner, respond faster, and spend more time on higher-value work. But sound design, clear governance, ethical use of information, and human accountability must guide it.
The organizations that benefit most from AI will not be those that adopt it fastest without direction. They will be the ones that combine technology with purpose, discipline, and professional judgment.
That is the opportunity in front of us: not simply to use AI, but to use it wisely.