As we’ve seen over the past year the security industry is charging ahead with the adoption of Artificial Intelligence, or AI for short, into security applications. There are several different use cases for this type of technology in security. In the video surveillance space is where we currently see the widest adoption. Technologies like Avigilon’s Appearance Search allow users to quickly search through video using Machine Learning techniques to identify People, Vehicles, and Objects as well as their motions and then rapidly provide feedback to the Search and allow it to re-run and refine the results. This process can be repeated to get incredibly high confidence returns in a matter of seconds. If you haven’t seen it before I recommend watching the following video: Appearance Search YouTube This type of use case speeds processes that traditionally would have taken dozens of man hours for an investigator to manually search and follow video to only a few seconds. This also allows users to identify patterns that they may not have known were there. We had an incident ourselves where we used this technology along with an employee’s report of a suspicious vehicle to find out the vehicle had been cruising our parking lot for a period of days. When we contacted law enforcement it made for a very different conversation and set of outcomes. This provides users with a counter-surveillance method that would have never been identified without the use of this technology. As the AI continues to learn and analytics continue to be refined we will start getting into a faster or even a preemptive response to incidents. This will be a game changer on many levels including the legal landscape where there currently are no legal precedents around applications of these types of technologies.
On the Access Control front we’ve seen moves towards more integration of systems, multi-factor authentication, and pattern recognition where AI is helping to identify possible threats, anomalies, and improve the overall security posture. One example we’ve seen in this space is using AI to recognize patterns of use by personnel and flag a possible insider threat. If someone is routinely early in and early out to avoid traffic, the system can learn these behaviors and identify when someone is changing this pattern like coming in late at night. This can be done in real time and similar to the video process doesn’t require monitoring logs and activities. The nature of the well publicized insider incidents over the past few years will continue to drive adoption of these types of technology into Access Control systems.
Overall we’re excited about the possibilities that AI brings to security, but make no mistake it will change the way we design, use, and think about these systems. We’re not far off from the Sci-Fi future of Person of Interest or Minority Report but we as a society need to take a lesson from Sci-Fi and think about how, when and why these technologies can be used and not just “if” they can be done.
By: Drew Weston
Drew Weston is the Director of Sales and Marketing for CodeLynx. Drew has been a technology evangelist with CodeLynx for the last seven years helping Federal and commercial clients understand the impacts to security programs from new technology and new ways of doing business. Drew has served in various leadership roles within CodeLynx include as the Facility Security Officer. Drew has an International Business degree from the College of Charleston. He is the Chapter chairperson for the Greater Charleston Chapter of ASIS and previously served as the treasurer for the chapter. Drew has been involved with NCMS, Infraguard and ASIS for the last seven years.