Like many people in the security industry, I recently attended ISC West in Las Vegas. And like many of those attendees I have some opinions on what I saw and what it could mean for the future of the security industry around AI products.
First, I would say that AI was matched only by VSaaS/Cloud as an overall theme at the show. And much like Cloud, the definition of AI was frequently vague or contorted. We need to work on refining some of these terms into more specific definitions around use cases or implementations, but that is another topic.
As it relates to physical security, AI has been primarily about removing burden from operators, and utilizing technology to help with tasks like searching through recorded video to find a particular person or object, or sending alerts when something happens inside a secured area. Concepts like smart search and video-based alerts have been around almost since the beginning of video surveillance, particularly as we moved to PC/Server-based software for storing and retrieving video. What AI has done is make these legacy concepts more accurate, more reliable, and thus more valuable.
It is worth noting that users don't really care HOW the results come about, just that they are reliable. If the user gets the results they want, then the product is good. If they don't get the right results, the product is bad. And of course these results need to be cost-effective to achieve. Thus, companies arguing about whose AI is more real, or where is the best place to run an AI algorithm are missing the mark on what users care about most. This is a trend I have seen continuously at ISC West, and I don't see it changing: AI companies are really really bad at marketing (with a few exceptions).