As technology evolves and threats become more sophisticated, traditional security measures are no longer sufficient enough to protect people and assets. Integrating artificial intelligence (AI) into existing security systems can help mitigate these new risks. While AI is a relatively new concept in security technology, its impact is profound and can deliver tremendous value for businesses.
At its core, security technology is a data collection system. Security has traditionally been a forensic tool, recording or tracking incidents for review after they happen. AI’s true power lies in its ability to convert this data into practical and usable information, a capability that security system manufacturers are actively developing. AI allows real-time data usage due to its ability to combine all components in a security system, process them quickly and report activities while they happen.
What is AI?
AI teaches computers to use more tools, consider probability and automate actions. It takes the if/then conditions in information processing and creates multiple levels of analysis and outputs; it combines millions of these processes in fractions of a second. AI ingests enormous amounts of data and creates useful information, making it a valuable threat assessment tool.
Unlike traditional security methods that rely heavily on human interpretation and intervention, AI can continuously process and learn from data. This means that AI systems can identify patterns and anomalies that would be difficult or even impossible for a human to detect, thus making AI the more effective option.
AI Integration in Security Systems
Intrusion Detection
Intrusion detection has traditionally provided immediate announcements when an event occurs. It uses the principle that “if” this condition occurs, “then” do this. For example, if a door position switch separates from the magnet in the door, then the device sends a signal to the system that takes that input and outputs an alarm condition. These events have been essentially linear, meaning one input event initiates a single output; one event equals one alarm. AI is transforming this approach by coupling multiple events, measuring probabilities to score the event and generating more intelligent information and responses. For example, AI can predict significant influxes of visitors at the main desk, which can help to deploy the necessary resources to manage that challenge.
AI can integrate with existing hardware, including door and window sensors and locks, motion sensors and more to improve the overall effectiveness of intrusion detection systems.
Identity Verification and Biometrics
Identity verification and subsequent access control have utilized AI in biometric analysis. Biometrics can take characteristics of fingerprints, facial features, iris scans and other biometrics, comparing real-time data to a database to identify a person within a predetermined probability level. AI enhances this process by expediting identity verification via biometrics, passively allowing or denying access, which helps ensure that only authorized individuals gain access to a facility.
AI-integrated biometric systems employ hardware such as fingerprint, face and eye scanners to capture and analyze data.
Video Surveillance
Video surveillance systems have historically only provided images for viewing and recording. Security staff would view real-time camera activity and past recordings to determine unusual activity. Video recordings have essentially been a forensic tool, and the time personnel need to review those recordings is extensive.
AI has significantly advanced the effectiveness of video surveillance because it focuses on making video systems more effective in real time. Whether the intelligence is located at the server or camera level, these systems can identify, track, and alarm based on color, shape, size, contrast, density and many other characteristics. Video systems can now immediately alert security personnel when the predetermined criteria are met.
Some video surveillance systems can use existing cameras paired with AI software on a central computer. Other systems incorporate AI into the cameras themselves.
Weapons Detection
Weapons detection system manufacturers utilize AI to improve these devices, moving from a simple metal detector to a more intelligent weapons detection system. The concept is to build algorithms that compute the probability of a person carrying a weapon. These systems compute density, shape and location data they receive as a person walks through a portal and compare it to an existing database of weapon characteristics. Based on a probability score, it will create an alarm event.
AI can integrate with typical weapons detection hardware, including radar sensors, cameras, X-rays and metal detectors.
Putting it All Together
As an example, motion detection devices can be coupled with video systems to reduce false alarms. Motion detection devices can be external devices that only detect motion. Not all motions require an alarm notification; some can be considered benign, such as moving trees or a bird flying. If you couple motion detection with AI in a surveillance camera that can discern between people, cars, trucks and animals, the AI will know to ignore any benign motion and not waste response resources. As far as AI in video, motion detection can be an internal attribute programming feature in a camera as part of that AI feature in visual surveillance.
System Customization with TEC
TEC is an independent consulting and design firm that is system agnostic. We are impartial to specific brands and focus solely on the most effective security system for a given client. TEC is deeply involved in technology assessment and can differentiate between real and functional AI versus false claims and unreliable technology.
There are manufacturers that market using the term “AI” because it is a broad and exciting concept, yet not well understood or defined. For example, some companies market AI-enhanced weapons detection systems that can “always” detect a weapon hidden under heavy clothing. There are features that these systems are taught to scan for, and they can raise an alarm (“red flag”), but these often require a person to conduct a secondary and more thorough investigation. Some systems require the weapon to be drawn and visible. These systems will do little, if anything, if the weapon is hidden at the screening point.
To achieve AI’s maximum value, TEC can provide the sophisticated design of multiple technologies into an integrated physical security program. It requires expert knowledge of AI capabilities on multiple platforms to design state-of-the-art security programs utilizing AI.
TEC Security Services
- Assess your technology and develop designs incorporating existing security technology to the greatest extent possible, providing a cost-effective and modern solution
- Provide AI-enhanced solutions that are either on-premise or in the cloud, based on system capabilities. More security systems are heading to Security as a Service (SaaS) solutions in the cloud
- Develop plans and designs that incorporate the integration of multiple technology solutions with AI into a solid security program
- Create plans that can incorporate future advancements in technology
- Assist as the trusted advisor and subject matter expert in assessing new technology
- Assist in assessing and training operational staff to ensure they understand and can use the technology
- Help facilities managers maximize the features and functionality of their existing security systems. Often, users are unaware of what their current system can do
- Help manage system programming to ensure that AI can be best harnessed and utilized
- Incorporate the security systems into “Smart Building” technology
Problems with AI in Security
In April 2024, ASIS noted that the primary challenge in embracing AI in physical security is the technology’s fragmentation. System manufacturers focus on implementing AI in their systems, yet the greatest value is the integration of AI across multiple systems. We are seeing advancements in integration, but this requires advanced engineering of these systems to achieve fully integrated AI across numerous platforms.
The Future of AI in Physical Security
Further Integration
Future developments in AI will be able to create an even more complex analysis of the threat landscape. Security professionals want to know their threats based on events, actions and who is present. While current AI is limited to each specific system, the future can be a network of disparate systems weaved into a threat analysis tool.
Advanced Real-Time Intelligence
In the future, security professionals may get messages sent to them that say something such as: “There is a 98% chance that the person that just entered through door #3 is the person that employee #205 on floor 5 has a restraining order against. There is an 87% chance he is armed. He is not allowed on premises.”
Conclusion
AI is still a developing concept, but it is undoubtedly growing and creating a more intelligent approach to assessing risk and threats in the physical security realm. It is still mainly limited to pieces of the overall security systems, so a technology designer with a deep understanding of AI capabilities, like TEC, is needed to fully harness the value of how and what can be integrated to create effective solutions. As threats become more sophisticated, so should our capabilities to mitigate them.
A nationally recognized safety, security, and emergency expert, Lauris Freidenfelds is the Vice President of Security Risk Consulting of industry-leader Telgian Engineering & Consulting. His 40+ years of experience include operational security, technology and emergency management programs. And, his extensive knowledge of emergency management personnel and operations encompasses planning, organizing, and directing security programs and activities, as well as development and coordination of disaster preparedness plans, the mitigation of, preparation for, and recovery from hazards and disasters. Reach Lauris at lfreidenfelds@telgian.com.