Sci-fi fans and Netflix junkies across the galaxy have welcomed the 2018 release of a rebooted Lost in Space featuring an all new cast, including a vastly upgraded robot sidekick for young Will Robinson. His signature warning, “Danger, Will Robinson,” has been immortalized in the annals of fictional space travel but really has a home in our own world of physical security.
The nameless robot is, of course, a robot, but what usually escapes mention is that it is clearly also possessed of many skills that look like what passes for AI today: vision, language, deductive reasoning, and more than a hint of cognitive abilities. All of these are, not coincidentally, a prescription for what we would like out of our future security systems.
The robot is a perfect security companion because it is totally silent except when there’s something truly worth remarking (also making it an ideal movie companion). That brand of non-distracting watchfulness is exactly what we want — or expect — of our security systems. Instead, we have noisy event streams that tell us far more than we need to know. And let’s face it: no one wants to look at their security system except when they need to.
Our problem today is many systems are not smart enough to tell us only what we need to know and leave us alone the rest of the time. That’s why we rely on dashboards to summarize complex data in a way our brains can readily assimilate, at least for now. If we had AI-powered event filters, like our friend the robot, we could go about our business like young Will, confident that the AI would let us know about any real threats or anomalies that required our immediate attention.
The robot also offers a voice interface that comprehends everything humans say — something just emerging in today’s security systems. There’s the handful of rudimentary Alexa skills for simple tasks like turning things on or off. But what we’d really like is the ability to ask analytical questions about past events, patterns, and anomalies. Even better, we’d like the robot’s ability to form wordless mental connections to humans, but I suppose we’ll have to wait a bit longer for that.
The robot also has a tremendous arsenal of facts at its disposal, as if it were permanently connected to the next-next-next generation internet. Add to that its advanced sensory capabilities (IoT, anyone?) plus communications with the cloud (interstellar version, no doubt), and you have a knowledge base that would be the envy of any Security Director. But it’s not just all this information that makes the robot so special. It’s the ability to apply AI to that information to keep everyone safe.
But the robot is also humanoid in shape and form. And that’s where Lost in Space diverges from the most useful security robots we’re likely to see anytime soon. The much ridiculed Knightscope robot fails precisely because it attempts to be an upright ambulatory intelligence, a mechanized security guard (with the unfortunate appearance of a Weeble).
The more useful robots we see today are like the successful AI: designed to be good at just one task, not to mimic generalized intelligence or generalized human behaviors. Think Roomba or autonomous vehicles. The successful security robots today do things like detonate suspected bomb threats, look under cars for explosives, and provide aerial reconnaissance and threat detection via drones. There’s no point (yet) in combining all of these into a single robot, and many reasons not to, such as cost and complexity.
In the same way that the new Lost in Space robot is much better than the old one (albeit not as inviting), our new robots and AI will continue to evolve. As they do, there will be more niche applications for them to accomplish specific tasks in the security workflow. What we’re all trying to figure out is what those workflows will be, and which ones will offer the best return.