PUT THE AI ON TOP
In explaining the role of AI in smart cities, a prominent scientist kept saying “you put the AI on top.” When I finally grasped what he meant, the phrase was so compelling that I tried it on others, it was like a light switching on in their minds, too. So I thought I’d try to convey that vision here, in the hope that it helps you.
Big Brother Was Already Watching
Let’s address the scary notion right up front: AI means Big Brother will be watching you. But this is nothing new: video surveillance is already everywhere you go and has been for years - in the garage at your office, the elevator as you go up, the halls, the park where you have lunch, and the store where you stop to grab a gallon of milk. Those cameras have been added over years in the interest of public safety and the protection of private or public property. And in most cases, they were installed after a preventable crime or calamity occurred.
So if Big Brother will be watching you, well, that’s because he’s been watching you for years. He’s simply been a terrible babysitter. Remember that time you accidentally set fire to the garage? It burned down because Big Brother didn’t know how to call 911. And that’s where A.I. comes in: we want that surveillance camera to see that fire and know it’s an emergency and do something about it.
Now let’s talk about the how. It takes about a million frames of fire footage to teach a computer to recognize an object or event. In some cases – logical emergencies like “fire” or “gun” – Google is indexing that footage and we can apply our algorithms to it. But there are going to be cases specific to your town, or your business, that are less common, and yet vital to you. In those cases, your footage is the data that will enable us to develop the AI feature. The more footage you have, the faster this can be achieved.
And if you’re on board with that, the first step to achieving that end is networking all your surveillance data together, so it’s all going to the same place.
Networking Your Surveillance Cameras
As we all know from making and storing videos on our phones, video footage eats up storage space like nothing else. The files are huge. So in order to work with that footage, we need somewhere with a LOT of storage: the public cloud is the logical home.
So how do we get that footage there? The same way it goes to a local server now – via a computer. Ideally, that computer should be dedicated to that task – a “Cloud Video Recorder” – because again, the files are big. The transport of the data has to be fast, and the processing capabilities for when it arrives need to be massive. And then there’s the cybersecurity piece: we don’t want anyone hacking into that footage and spying or tampering with your data, so that portal needs to be safe.
Once you have that data going to the cloud, you can “Put the AI on Top.”
The AI/ML Supercomputer
Think of Machine Learning and Artificial Intelligence as a giant Supercomputer in the Cloud. As mentioned, it needs a million frames of dedicated footage to teach that supercomputer. And because the files are so big, we need a lot of processing power and bandwidth to process it. Engineers have those things in the cloud. They can funnel all that footage into their AI algorithms, and teach the computer something new, so it can understand not only what it's seeing, but what to do when it sees that object, person, or event.
In our example, the supercomputer learns to call the fire department. It can even be taught to send Mom and Dad an alert so they race out of the movie, drive home, and make sure you and Big Brother are safe. Even if you’re grounded forever for starting the fire, at least you’re alive.
An “Eye” to the Future of AI
Artificial Intelligence and Machine Learning are still very young. There’s a lot we have to teach the computer, and a lot of things engineers have to figure out how to teach the computer. But the possible applications are endless. Eventually the police could conceivably find a kidnapper’s car in real time, without having to crowdsource an Amber Alert. An ambulance could be sent for someone who has collapsed in a company basement, alone.
Will it need ethics, especially around the preservation of privacy? Absolutely. But we didn’t deny ourselves the wheel because some accidents might happen. We built the wheel for wagons, and then trains, and then cars. The rules guiding safe operation of each evolved out of their use.
There’s a lot to do, but if you’re interested in smart surveillance, rather than throwing your footage in the trash after 30/60/90 days because there isn’t room on the local server to keep it, it’s time to talk to a cloud partner.
Here at Cloudastructure, we’ve embraced the future of AI Surveillance like a bunch of astronauts boarding a rocket ship labeled “Final Destination Unknown.” We’re ready and excited to meet you and hear about your challenges, whether it’s smart security or a smart city, and we’re going after the software piece alone. You pick whatever hardware you want. What’s important to us is helping you “Put the AI Top” today, so you’re ready for tomorrow.