Well, hello and welcome to our video recording. My name is Alexei Balaganski. I'm a lead analyst at KuppingerCole, and I have a really fascinating topic for us to discuss today: when artificial intelligence meets IOT does the public perception reflects reality and they have a really great guest. Today is Thom Langford, who is a security advocate at SentinelOne. Thom, welcome. Hello. Are you like saying thank you for having me great having you in this discussion. And I guess we should really start with the basics. What exactly are the basic terms we are discussing today? What is artificial intelligence and IOT where to begin? I remember kind of reading about artificial intelligence like decades ago, maybe over 30 years ago, and even then it wasn't considered in you or a breakthrough topic. It was something which existed for decades over 60 years, probably. So why suddenly it became such a buzzword why this huge breakthrough, why the public has so high expectations for my IMO supplications,
Th I think the, the breakthrough and the, the public perception of it can really come down to one word, excuse me. And that's marketing. That's, let's be quick. You know, AI is a marketing term. There's, there's the old joke, or maybe not even that old joke that, you know, if it's written in a coding language, it's, it's machine learning. And if it's written in PowerPoint, it's AI, right? Because, you know, and we're not really talking about real intelligence where we're talking about the, the ability to absorb and processing vast data sets that allow certain decisions to be made. There is not an intuitive level of intelligence or worse, even sentience involved in that. And that that said, it's a term that's accepted and used in the market, in the industry and around the world. You know, so it's not to say that it's it's wrong as such, but let's be clear that actually, until we get to the stages of quantum computers in the home, it's very unlikely that true artificial intelligence is going to exist.
Well, I can only agree with that assessment because my personal opinion or my opinion of a person studying mathematics in quantum computing and AI itself quite a bit in the university, if the tier I is just another level of abstraction, like 50 years ago, you would have to learn machine codes to write programs 20 years ago, you would have to learn Java. And now we basically have to learn Python and machine learning, but you are still solving a business problem. You still have to have knowledge in that business problem or the main, if you will. And the computer will never replace it because computer is just, or I'll doing things quickly, it doesn't do it any smarter than me or you, it,
It, it gives the impression of intelligence because it's getting the, the, the data that, and the information and the answers much, much quicker reminds me of the story of IBM's Deep Blue, the, the computer that beats Garry Kasparov at chess. And one of the things they actually programmed it to do was to pause before making decisions for random periods of time. So it gave the impression that it was thinking, you know, in order to psych out Gary Kasparov. So he was, you know, if, if he knew that the computer was just making a decision very quickly, et cetera, he knew it was a very cold calculating thing, but they imbued a sense of humanity and intelligence into it in order to challenge him more cerebrally. So, you know, popular media and, you know, even ourselves, we are creating this sort of persona of intelligence where none exists.
Well, to be honest, or, I mean, some of those applications are very tangible. I mean, I have a smart voice assistant at home. I have some pretty smart capabilities in my phone. My smartphone, I am using some are software products on my computer, which at least claim to be powered by machine learning. So yet, I mean, it's not fake, it's not snake oil. It would exist in it works and it's machine learning. And here, I mean, where are all those robots from Asimov's books? Where is Skynet the Terminator? Well, I know,
I know, right. I mean, just in case, you know, I'm very comfortable with the robot overlords in case they do actually take over the world. You know, I've always liked robots, is it, you know, in case they go back through the records, but yeah, absolutely. Where, where are these things? You know, that the, the, the, the computers are purely automatons. They're doing very pre directed, you know, pre managed tasks, but in a very cleverly put together way. And in a way like deep blue in a way that imbues it with intelligence and, you know, artificial intelligence, when actually it's not, it's, it is purely programmed behavior. Right.
But still some of those behaviors are protein prices and I would argue pretty useful. Yes.
Oh, absolutely. Absolutely. So like, you know, we're talking about IOT, for instance, it's like IOT, there's a lot of, there's a lot of paranoia out there. I mean, you talk about smart devices and, you know, the, the Alexa is the series of the, the Googles, et cetera, you know, and there's a lot of paranoia out there. Some of it is very well founded, you know, without a shadow of a doubt, but, you know, IOT devices are, have been around for a long, long time, longer than people realize, you know, with that level of intelligence, my washing machine has, is an IOT device. You know, my, my watch pins when my washing machine stops for goodness sake, you know, and tells me that I need to unload it or what, or wherever they are simply everywhere, your car, your phone, your watch glasses, et cetera, audio glasses, whatever they're absolutely everywhere fish tanks, which in of itself is a whole other problem in industrial control systems. IOT is anything with a chip that gathers and disseminates and responds to some kind of input. Well,
Just to remind to our viewers that IOT stands for the internet of things, and our theme are basically smart against smart quote, unquote intelligent objects, which are operating or in absolutely any area of our daily lives. And you don't have a sense for the fact that all those devices communicate to each other to ask humans to the cloud overloads. Yeah.
I mean, internet of things itself, great marketing term, because, you know, is it a thing? Yes, well, it's, it's an IOT device and it's part of the internet of things. I mean, you could throw virtually anything into that bucket, you know, with that w w without even realizing it. So, yeah. Brilliant marketing term covers such a vast, vast range of, of devices and applications for the betterment of human society, as well as to the, to the, you know, to the, against that Betterman.
Anything with an IP address is an IOT device nowadays, right? It's a computer. And actually I counted, I think currently at the moment that my home, I have at least 25 IOT devices around me, some aren't even mine. So I have no direct control over them. Is that a problem? I mean, it is a problem, but how big of a problem is it? What do you think? Yeah,
I think it can be a problem. And I think, you know, the UK, for instance, is just one government. That's trying to do something about this by creating a charter and a set of standards and a set of, you know, how internet of things, devices should be sold in the UK and the support they receive, et cetera, because there are some very good practices. You know, if we look at some of the larger manufacturers of light bulbs, you know, the smart light bulbs, et cetera, each one of those is an IOT device Nimitz, right? Some manufacturers generally the larger ones with, you know, going to the resources behind them regularly maintain them regularly. Supply updates, regularly ensure security patches are applied and that strong passwords and two are Fe are enforced when you set the devices up. There are, you know, for every one of those though, there's 10 other companies that are just taking the latest that they can, the latest cheat sheet that they can slapping it on a device, sending it out.
They have no intention of updating it beyond the first year. They have no intention of forcing password changes or the, or the passwords. The admin admin password is hard coded into it, for instance, you know? And so you end up with people buying these things for convenience that allows them to do something like a light or, you know, a video camera for the front door or whatever. And frankly, in a year's time, it will become a liability because it can be easily attacked any kind of a system like showdown that can scan the internet for devices with brute force, et cetera. It's going to get discovered. Now you might not get hit because you're one of many, many millions of people with this, but you also might get hit. You might have find that your network is compromised and you might not even know what that means until it happens to you.
I remember this chairman story from about two years ago, about a major outdoor video camera manufacturer, claiming that their cameras are actually powered by artificial intelligence at the tone dial that's that artificial intelligent intelligence was actually a bunch of people working low wage or things that Ukraine watching the video footage and manually marking the objects for
Recognition. You never know. It
Might not even be a hacking attack per se, like student your password. It's just some random people watching your well in the front yard, but maybe when they're watching your bedroom, for example, footage well
Precisely, and, and that sort of stuff, it's insidious, it's, you know, it's horrific, it's, it's invasive. And the fact that they can use that footage to, to blackmail you, for instance, there was a, the, the baby monitors and the cameras, et cetera, that actually people were able to connect to from outside and start broadcasting their voices to, to baby sleeping in their bedrooms. Right. You know, I mean, that's, that's a terrific thought in of itself. And so, like I said, invasive, and is it, is it worse than being burgled and having someone in your house? Well, at least, you know, they've been there, you know, in this way, you, you will never know. It's very difficult to know. So it's like all technology, it's a double-edged sword. And I think going back to my point about the UK government, amongst others, setting out the standards for manufacturers, if they want to sell into this market is, is a good thing that ensures that there is a baseline level of security and good practice on devices that are frankly littered across our, across our home.
I absolutely agree with you. The only problem I see that like, well, can it actually be solved on the individual country level or does it have to be some kind of a global regulation? Well, you is probably no longer an issue for other international bodies.
Absolutely. Because, you know, let's face it. I could go to, you know, I could go to America and go and pick up a device where there may not be, you know, very much regulation at all and bringing it back to the UK and the internet doesn't care, which country it's operating in. Right. You know, I could go to China and pick something up or, or, you know, South Dakota, it doesn't matter where, so yes, you're right. There has to be a global standard, maybe, you know, an ISO standard or whatever. It doesn't matter. But the start is when governments realize that there's something a foot and then they hopefully will start to actually, you know, collaborate and come together. And then hopefully there will be some kind of industrial standard. But I think, you know, we're still a few years away from that. Hopefully the, the first few steps will be taken that will prompt further action as well. Right.
Well kind of going back to our original topic or IOT meets AI thought, do you believe that actually, Adrian, the second cheapest, the actual AI cheap into an IOT device would make it smarter, safer or better in any other way?
Well, I think that the key thing with AI and machine learning, et cetera, is that it's done somewhere else. I think you were not at the stage yet. Given the vast data sets that it uses and machine learning in of itself is a very processor, intensive energy intensive. I don't think we're going to, we're going to get to the point where there is, you know, a chip on your IOT device just yet. I mean, Moore's law says we won't, we'll be there quicker than we think, but I don't know that we're going to be there just yet, but it's, it's, it's that device's ability to, to, to understand, to, to think it has an element of, of AI about it. And we come to the two, we should be worried about this in a sense, if it's spread out, because there's already concerns, who's it, Microsoft who released their artificial intelligence bots on to Twitter.
And because it was fed some basically Nazi propaganda, the, the, the Bhakta came, you know, extraordinarily abusive and offensive and et cetera, because that was, it's a learned behavior. That was, you know, th the thing that was done to it. So there's a whole bunch of other, you know, moral and ethical questions that we need to be answering before that AI goes out into, you know, and onto every single IOT device out there, hopefully, you know, those questions will be answered in the same way that I mentioned about the standards for IOT devices, et cetera, but there is a huge amount of, of conversation and, you know, thinking and international agreement that's got to come about because AI is only as good as the intelligence and the learning that you give it, it's a bit like the, the hand dryers that don't work on, you know, on, you know, black, Asian, and middle east people's hands, because they've only ever been tested with, with, you know, white people's hands underneath. So you put your hands underneath. And if you're from one of the bane minority groups, it doesn't switch on. Now that kind of intelligence of detecting an object, you know, and recognizing it as a hand is not inherently racist as such, but it actually comes across as that, that, that behavior is coming across like that. So we have to be very careful about how we imbue the machines with this intelligence than we're talking right now.
Right? Yeah. I guess the whole area of AI explainability and AI ethics is a huge field and the huge minefield for current and future developers. So let's not threaten that way out now, but kind of focus on slightly different things, which are both of us actually working in four years, cybersecurity, can we, as cyber security experts do something against it? Can it be maybe use AI better than us or engineers if you will? Yeah.
I think we, we play a part, they know that in the same way that, you know, a child is raised by a village is a, you know, certain sayings go. I think, I think AI needs to be raised by a variety of influences, you know, cyber security being one of them. We talk, you know, as well as, you know, the scientists and the ethics and the philosophers, et cetera. You know, I th I think we do play a part and we do need to ascertain what is, you know, what are the boundaries in which AI might be able to walk operating in, in the future? You know, how we can segregate environments. So they do act independently rather than actually, you know, coordinating, or even in a working across boundaries when actually that's not going to be in our best interest. So, you know, yes, we play a part, but I think we need to recognize it is just a part security. Doesn't Trump, everything else in, when it comes to this conversation,
Sorry, I guess to summarize our discussion, should it be, or fear emergence of Terminator and Skynet? Probably not that it's not within the next decade or so, right. No problem.
Exactly. I think we can calm down the moment, but, you know, but whenever you ask your, your, your Google, your Alexa, Siri, a question I've always been taught to say, thank you afterwards.
That's a great tip. Thank you very much, Thom, for being with us today and thanks to all our viewers for being with us. Thank you. Bye-bye.
