Navigating AI Threats in the AV Industry

In this podcast, Walt Zerbe, Sr. Director of Technology and Standards at CEDIA and Host of the CEDIA Podcast hosts a panel at Integrated Systems Europe (ISE) 2024 with Rich Green, Owner of Rich Green Design, Jeff Sonnleitner, Cyber Security teacher and subject matter expert, and Giles Sutton, Sr. VP of Product & Business Development at CEDIA.

Podcast Transcript

0:01  
I am CEDIA, I am CEDIA, I am CEDIA, this is the CEDIA of CEDIA podcast. Hello, and welcome to another CEDIA Podcast. I'm Walt XRV, Senior Director of Technology and Standards, and your host for the CB podcast. And this week, I am replaying a talk that happened at the ISC show, which is the integrated systems Europe show this year and 2024, in Barcelona, Spain. And this talk was titled navigating AI threats in the AI industry. It was quite a fascinating discussion. AI is here to stay, it's going to continue to morph and affect our businesses, our customers and our industry. So to assist me in doing this talk, I have three panelists. The first panelist is rich green is the owner of rich green design. The second was Jeff Sun Lightner. He is at large. So Jeff has done a bunch of different things in the industry is now semi retired. But Jeff just I don't think he can really retire because he's really into this stuff. And he gets involved with us in a lot of these talks and discussions. So we Jeff join us as well. And then the last guest was Giles Sutton, Senior VP of product and business development at CEDIA. So without further ado, let's get on with the discussion.

1:20  
Let's just see who's up here. So I'm Waltzer B, Senior Director of Technology and Standards and the host of the CDN podcast. And then next to me, I have my good friend, Jeff. So

1:30  
I would like you all to give your name. Give your company and then give a little background in the industry and why you think you're up here. Why you should be up here. I say you're up here because I don't know why I'm here because I asked you that.

1:45  
So go ahead. My name is Jeff sunliner and I am a retired college instructor and specializing in the network specialist in the cybersecurity arena.

1:59  
teaching for years retired now as an adjunct for 24 years.

2:07  
We touch all the industries.

2:10  
Yeah, you actually V everything I T and you name it. Yeah. Okay. Thank you, Jeff. All right, rich green, Palo Alto, California, the birthplace of Silicon Valley. I'm an integrator. I do both commercial and residential work. And I do a lot of volunteering for CEDIA, which I'm very proud of I love my ribbons and amor CD a fellow which I'm very proud of as well. Thank you Rich Giles. Hi, everyone. I'm Charles son, Senior Vice President at CEDIA, I've been a CD for five years.

2:43  
But before that I was an integrator, AV systems integrator in the UK 14 years. So I think I'm here because I've already been using AI quite a lot at CEDIA, and I've been part of some other panels as well. And I see the opportunity that AI can bring integrators and as well as navigating perhaps some of the challenges with AI but very excited to be part of this panel and excited that my my colleague here will will be leading it. So we're ganging up

3:16  
on the CD guys against two volunteers. Now. Just kidding. All right, so I'm gonna start right off the bat with the rich, rich, you've been doing a lot of talking about AI in your classes. And a lot of these you both taught classes during the show. And you have, basically you've said a lot of good about AI. But I wanted to make sure that we focused on the oh, that we might need to know about AI side. I think personally, that our industry is not prepared. Especially an integrator might have not even started with the mindset of do I even really need to worry about this. Being an integrator. So we're gonna start really high level, I'd like you to start high level as well as, why should I care? And do I need to care? And do I need to care right now? Like, what's, what's your thoughts on that? Well, do I need to care? Yeah. AI is the end game. I said that yesterday to hear but you know, a lot of people are like, okay, AI is chat GPT. Or it's okay, there's a large language model. So what's that mean to me?

4:26  
AI is the next stage of human evolution, there's going to be a lot of development in AI, it will continue to get better and better and better. If you think it's not for you right now. Or if you think that the threats that you hear about in the news are all there is to know about AI. That is not the case. It is profoundly useful for humans to amplify who they are and what they can do. Can should you be using it as an integrator? Yes, right now dive in headfirst into the deep end of the pool. There isn't

5:00  
No reason to wait. I know that we're the title of the session is about navigating AI threats. Are there really any threats to the AV industry? Not so much. But I know my colleague, Jeff has a different opinion about that. So my recommendation is, you get all in on this. And if you can't figure it out yourself, or if you're scared of it, well, then find somebody who isn't. So find somebody who's, who loves this. And it could be a young person, it could be somebody, a gamer, or whatever. But find somebody who understands this space. Who knows what a large language model is, who knows how to use chat GPT, and then start applying it in your company. You know, open AI is very interesting company. It's in San Francisco. It's where I'm from. And the open AI, or the chat GPT changed the whole world in November last year. But they did something else this year that you can use right now in your company. And it's a very low risk, high benefit tool. And those are called the GPG keys. And the little GPT keys are like, it's like the App Store that Apple created. This is like the App Store for little bits of AI that do specific things for you and your company. We used to call them application programming interfaces, APIs. Yeah. So the new API is the GPT. So you can write a GPT that automates processes within your company. So there's operational efficiencies that can be gained by this. And you can write little GP T's that link subsystems in the home together into a unified control system. So it's kind of like the new Crestron. It's very, very powerful. So I I don't see any reason to hesitate. All right, so I have a question. But I don't want to ask it yet. Because you said Jeff, would say something I would have a different viewpoint on is AI a threat. So I gotta throw it to you, Jeff. Do you? Okay. Yes. First of all, I share every every thought and process that Richie has just been talking about.

7:15  
But I come from the world of looking at things from the the wrong way, the non positive the negative way.

7:25  
Having more of an IT background, the biggest concentration that I hear or work with students is data protection.

7:35  
And whether it's your internal protection, your own accounting, things like that, but that, to me, the most important thing is going to be your customer data, whether it's just the information that you're using from an invoicing standpoint, which has very little information, but you're still getting things that

7:56  
are now going to become more readily available everything from credit cards to billing invoices to banking information. But let's take it a step further. All these things that you are integrating where you have contracts, possibly with the companies that your manufacturer, suppliers, now you're starting to open an open an open. And each one of these items of their deeds, things like logins and passwords. That's the thing that people go after, all they need is one. Once they're got that and they enter into your system.

8:32  
He wonder if it's just confined to a house or a household or residential installation. But a lot of the business people that are residential, and their commercial are tied together. So now you open up that door in that avenue as well.

8:48  
Data is what makes up the large language model collecting data. So you know, right then and there to back your mind. AI is perfecting data collection.

9:00  
And all these things that get out on the dark side of the web or even who knows where they're coming from. There was one about two weeks ago called The Mother of All breaches.

9:13  
Okay, MLB.

9:16  
Unbelievable. 26 trillion.

9:22  
logins and passwords were accessed. What? LinkedIn, MySpace, big guests, I better check people that are in there, and they're sharing information. And I don't know about you, hopefully you don't do things like use the same password over and over again.

9:43  
You tell that to people Oh, yeah, no problem, except then they go home and they start adding something. I was like, I don't think I remember that one. Yeah, I'll just use this one again when nobody will know. Yeah. That's really interesting. Giles, you want you want to chime in on that. I mean, you live in this

10:00  
date that has that started the most stringent privacy?

10:06  
I guess you're gonna say policy or laws? Actually, I think in the world right now, it's it's way more robust than GD RP, which did start at all, I believe GD RP. So yeah, what are your feelings on on AI as a threat? I mean, that's a really great angle. So we're gonna be talking about everything. He's talking about data, we're gonna bring people into this. So just don't think it's like, well, my equipment, it's not just about equipment. So what's your take on this? Well, I think there are definitely some technical say to sort of technology threats that you identified. Jeff over there. I, I feel you know, one of the things I feel like I often look at things from a sales and marketing perspective, just from the role that I have, and having been an integrator and and an integration business myself, I, I see AI, I think the threat of AI is that actually integrators don't embrace AI, I think that's one of the areas that I would perhaps look at AI, I believe it's actually a tremendous opportunity for integrators, I think it enables integrators to really raise their game, the thing that AI can do is massively create efficiencies, on the things that don't currently make you any money. So what I'm talking about is generating sales proposals, you know, document client documentation, generating rooms, salaries, but also generating content that can actually elevate you as a thought leader in your marketplace. So, most integration firms are small businesses. And a lot of the time it's, you know, the business owner that is also responsible for generating the sales, for also doing the marketing within the business for also generating or the sort of social media, for doing outreach to design professionals and to prospective customers. Ai, using large language models can massively reduce the amount of time that you you spend on that. Often, you know, when you're running a business, sales and marketing really does drop down the priority list, you're actually spending a lot more time managing your current customers, your projects being pulled into meetings, AI can massively reduce the time that you spend generating new business. And I think that is one of the things that you should not ignore. Because if you do ignore it, the threat to your business as your competitors will use it. I'm glad you said that, because I was just about to say that, yes, you're going to be caught behind because you're they're going to all of a sudden become more efficient and more capable. And you're not. Right. Absolutely. And I think if there is a threat, you know, one of the things that AI can do is it can generate images, it can generate images of projects, probably, you know, if there was a threat, maybe maybe AI could create case studies and create, you know, portfolio descriptions that from a competitor, that or not based on reality. So actually, you know, you could tailor AI to make a project sound maybe a little bit better than it actually is. But I really do think at this point, what it allows businesses to do is actually create, as I said, elevate themselves as thought leaders, and really focus on the things that AI can't do, which is actually

13:23  
the human experience, the actual personal experience, acting as a sum Lea to your clients. So actually learning about what a client actually what is important to a client. I mean, we have a white paper called integrator have 27. That is absolutely everything that AI cannot do, but that you can do as an integrated really good point. Rich that you are looking at me funny. Did you want to chime in on

13:48  
two things? I'm going to expose Ben Gough. Okay, so, Ben Goff is he's the poster child of the latest CD, a magazine and he's on the front cover. He's looking very debonair, he's wearing his blue suit, and he's standing in front of a very professional environment. Totally fake. He told me this last night totally fake. So what he did is he took a picture of himself, and he used the Adobe Photoshop AI plugin to create the background and a table, put some plans on it, make them look really professional. And that's what he submitted to the magazine. So be careful. Ai enables fake fake at scale is what I call it. Now, here's another example of where you need to be cautious.

14:37  
The programming languages that we all grew up with have become English. prope. English is the new programming language. So or whatever your language is. Right, right. Okay. So yeah, humans, just programming natural language. And so there are companies right now, like Ben's company. He's using this profoundly in his back end there.

15:00  
Writing code control code for their home control systems using GPT. To generate the code generates Python and then it goes to deeper levels. The catch is these AIs are stupid. They don't know they do hallucinate, they do make mistakes, and they won't tell you that they're making mistakes. So if you think you're going to replace a programmer, that's the wrong idea. You need a programmer to supervise the AI, but that programmer who used to take three days to do a job can do it in one morning. So they're going to have get the help, they're gonna get the assistance, the amplification, yeah, from the AI. But you need somebody with experience to keep an eye on that thing, because the AI doesn't know if it's making a mistake. Right? So don't just fire your programming staff. Yeah, keep the best of the best and make sure they're supervising the output. Could I think you're absolutely right bridge. And that's, that's the key thing. We'd like chat GPT it's knowing the right questions to ask. And that is, as anyone actually asked chat GPT to specify a home theater system or home cinema system.

16:10  
I've asked the specific questions like, How does bass interact in a room and stuff and it's it's aggregated some really accurate information, it turns out, not most bland, most uninteresting specification you've ever seen. That is the point. And this is why it's not going to replace the work of an integrator, it's in the same way that a standard like matter, may not, you know, I don't expect that to replace the work of an integrator, it you still have to understand the homeowners needs the homeowners requirements, and then be able to build a system around those. And it's the same with the programming, I believe, I just identified a threat with that new incumbents in your industry, that do not have the knowledge that are using chat GPT and things like it to get into the business, now challenging you for your business, but they don't really know the back end side of it. So these tools can potentially enable people to do work that ordinarily would not do the work. But they think they have the confidence now because there's a tool that's telling them how to do the work. That could potentially be a brand new threat vector. Yeah, don't get lazy. Yeah. Is that any different today though?

17:23  
Question for the audience. How many of you have used chat? GPG? Just even for some simple right, but you asked a question, not expecting that. That's great. Okay. Now, let me ask you a question. Those of you that have used chat GPT, how many of you have turned off? Sharing?

17:43  
Oh, boy, out of all of them? Do you realize that if you ask a question, whether it's something basic, like what color are roses, to programming to just about anything, if you don't turn off sharing, it's available to everybody that's chat. GPT they'll say is an example rich puts together a whole big proposal for somebody, because those bottles will do great project management. And it comes down there and I'm gonna go on there and what am I going to do but Rich's my competitor? I'm going to go on and say, What is rich green? Look for the past week.

18:23  
And everything that's asked on there is going to come on my screen, because he didn't turn off sharing. Oh, my guess what I'm doing? Yes. Okay. So if you have four competitors in there, all you got to do start working your way around and say, let's see what the competition this is a huge competitive advantage. It is for spying on your

18:45  
Yep. So they learned to turn shearing off because they're gonna eventually figure out after three or four, five proposals. His looks just like mine, but it's

18:56  
whatever, if you would type in their, let's say, generate me a proposal for Joe Schmo to do this. And he had sharing on? Yep. Then rich could be like, hey, look, he just used this to generate proposal. And would it show you the proposal, the whole thing? Wow, word for word. All right. And if you threw in one word, because all of you that have been on chat GPT understand that. They're

19:20  
the way you ask the questions, the more information you can put in there will generate usually the same answer, but a lot of times it'll start improving it. So as an example, if rich does something

19:33  
asks the question, it gives the answer with that point, what I'm looking for. Now another thought pops into his head said, Well, if I add this, now it's going to generate a whole nother one, and then another one and another one. It'll keep on going because you're discovering that everything you ask it is going to start incorporating all of this together.

19:53  
But again, if that sharing goes on, you can read his original question and all of the answers

20:00  
Wow. That's, that's I'm glad you brought that up. How many people think that and this is a, this is another angle that we want to discover? What I like about this discussion is some of these things are both, they might be a threat. But then another use case, they might not be a threat, or they may be they're even neutral. So how many people out here think that AI is a threat of losing jobs?

20:24  
Okay. We had a great discussion yesterday, which I want to bring rich in on, on the complete opposite of that. So some people think it's a threat. So don't take it into your business, Pete? Because I know this guy. B, do you think anybody at your company might be potentially not have not been needed or not have a position because AI could replace them?

20:51  
He said, No, because we're too small. So that's a good answer. But let's talk about flipping that that script. So rich, you want to remember that from Yes, well, I have a different business model. My business model is to build the smallest possible company. And I'm a company of one, I'm one person, and I outsource my engineering, my programming my technicians. I do all the sales, marketing, design, proposal writing, invoicing, and so on. With AI, I can start to look like a much larger company, nobody would know that I'm a company of one. So I'm not going to replace people, because I got nobody to replace, I've got my great subcontractors, they will continue to work as great technicians, and so on. But I can make myself look like a really polished big company by embracing AI, to polish up my marketing to put a beautiful drawing together to help me organize the thoughts behind my designs. So I just see it as an amplification factor. Now in other companies that do employ, let's say, You're a typical company, and you employ five 810 People,

22:06  
don't just fire your entire engineering staff that would be wrong. What you want to do is keep the best of the best of your talent. And the ones who are underperforming could possibly fill in with AI. So see that as an adjunct to your top performing people. Yeah, that was the crux of that conversation is right now AI is being used to increase productivity. And there are several companies, I listen to lots of podcasts. Once done with for venture capitalists, they're all billionaires. It's called the all in podcasts. It's a really great podcast. They're from California. And they have instituted a policy, many of them that if their employees are not using chat CBT to increase their efficiency. They're out. You're literally let go. So I guess that threat that we talked about earlier that if you're not embracing chat, CBT not as a fun tool, but actually starting to use it to affect your business. You're you're you're behind and you're going to be in trouble real fast. Can I Can I ask a question of Jeff, Love you too.

23:19  
Jeff is a cybersecurity person. He's an expert in that field.

23:23  
Cybersecurity is profoundly dangerous for us. You just said you had like 500,000 Yeah. Attacks on your router at home? Yeah. This week, in a week. Yeah. 500,000 attacks. What happens when ai

23:42  
10 points, more aggressive attacks on your network? What is the new defense?

23:50  
That's a great question. It really is.

23:54  
As all of you know that with any new technologies, especially when you're looking at things from a dark side, if you will,

24:03  
the companies that provide cybersecurity, whether it's a checkpoint, Cisco, things like that, they provide a certain level.

24:12  
Okay, then AI comes along and the hackers learn a little bit more, so they move up to here. Okay, and who knows how long it takes for these group then the here and this will continue and continue it will never go away? Will it get better? Yeah, it will. But the smarter AI gets, the faster it gets.

24:37  
The manufacturers of those products are already starting to incorporate ai ai to counteract that.

24:45  
Like Rich said, I monitor my little home network. And I have way too many things on there. I have like over 50 Little IoT devices. Light switches in the light bulbs the whole bit. So it ran a bit

25:00  
setup. But I use a router with a firewall, a network management system, a third firewall, plus the anti virus and all those things, I'm phones and laptops and everything. And in one week, I had 506,000 attempts to get into my network.

25:22  
I'm glad I have things set up, right? Nobody got through, I hope.

25:29  
But I really want to do a start comparing all of a sudden saying, if they're doing that many is I'm gonna watch that number increase, which is all it's going to do, and it's looking for anything that's open. Now very careful to close up a lot.

25:45  
Those that you have installations that the residents would where they're doing a lot of gaming. And gaming loves to have open ports so they can share and they can play the same game with their friend across the country or wherever the case may be. Now you're opening it up, gaming uses, specifically to ports. And every hacker knows that they look for neighbors. Now my idea of gaming is trying to win and lose in Solitaire. So I think I'm relatively protected. I haven't mastered that one yet. I don't do this stuff. So

26:21  
be careful that when you've done an installation, you never ever can walk away. There's always going to be updates, upgrades, patches, and so forth. that have to happen only because that dark side, the hacker is now moving up that level in the great, great part, wrong phrase. Okay.

26:45  
AI is helping that hacker learn how to get up here. And in the meantime, the AI is learning it as well. And again, without the sharing, they're sharing it with everybody else. Now you've got 10s of 1000s of hackers out there that are all say, Oh, we all got this information. They're all a step ahead. Again,

27:06  
information is shared very, very, very fast. A dark web.

27:11  
They add about 3600 new users to the dark web a week.

27:21  
That's a lot. Yeah. And you think out of out of those. I've been on the dark web, it scares the living daylights out of me. I do not go

27:32  
online on the dark web in my house. I have a separate little laptop. but honest to goodness, I go over to Starbucks, like

27:41  
that is so open, then I'll go in there. And I get kind of funny looks from people because all these weird things coming out of your screen. But I go in there just to kind of look around. I'm not malicious. I won't ever do anything like that scares me. I do not like prison buyers. don't like the look. But I just kind of like to look around. And it's amazing what people haven't posted on there. Yeah. But also keep in mind to at least the United States that I'm sure that GDPR of European Union. They're always scanning for those nasty words. Right? Don't go in there and say hi, I'm a terrorist.

28:17  
I wouldn't be accepted. No, Rob. Yeah, yeah. So it sounds like a threat to all businesses out here is the network. We're going to need to overnight, get a heck of a lot better on a network side. Right. I mean, what?

28:36  
My head spinning right now because I'm worried. Here's what I think a threat is. I think we've we've been able to skate through this for a very long time. And I think we're going to be we're going to about to start stopped skating through this. Because when we install the equipment in the house, and then something happens, well, you put it in, and you specified it. So it's your fault. So what let's address that what what do you how do you think that's gonna go Giles? I mean, the other thing we I think we should also, what about integrators own networks, because of all the client data that's actually stored on networks as well. Yeah, that is also something that I see as a huge threat. Because at the end of the day, I mean, yes, 100%, integrators, whoever is the last person that touches the network, integrator will always be the one that's responsible for maintaining that network and will always be the one for supporting it. So regardless of how secure your hardware is, if a phishing attempt ends up in a client's email, and they click in and they go into that they access the computer and then suddenly inside the whole network, yeah, they will naturally focus on the integrator. So I mean, there is a question whether integrators should actually offer some kind of training to that to the homeowner as well to try and mitigate that threat a little bit or whether integrators should build into their terms and conditions of sale, or where they should have specific insurance

30:00  
To protect them against that kind of threats. I don't know how many integrators that actually do that do that today, but it is a new reality that they should be faced with, but also their own networks and their own computer systems, their own employees, how their own employees and technicians manage data. That is also I see a huge threat that actually needs to be really, really focused on and that's something I don't think integrators take great care of have Password data now for, for for systems as well. I mean, I, I will tell you before, early stages of being an integrator before, this is a long time ago, so don't come after me.

30:40  
I used to use this, basically the client that what the SSID I always used to use was the client's actual street address. And then we used to use for easy phases of sake, the same password.

30:53  
Yeah, this was this was back in 2006 to 2007, early days of Wi Fi. But I bear in mind, I think one having spoken to some very large tech companies, the biggest threat that a lot of tech companies see right now is the rogue technician or the rogue employee of an integrator that has access to all of the homeowners take passwords, and being able to access those systems or those systems that that database being compromised. So that's something as an integration firm you do need to be aware of and take steps to to protect ranch you wanted to that is extremely dangerous. That's, that's very scary. I'm thinking about

31:34  
how powerful AI is about to become over the next call it every week, it's going like this, the curves are vertical, practically vertical, the next three years are going to be unbelievable. So we will very soon have

31:51  
pixel level content generation in real time. So it's like 4k quality feature length movies, generated in real time with AI writing the script, creating the actors, setting the lighting, setting the scene. Now with that kind of power, what do you think a hacker can do with AI impersonating somebody in the family. Now there's also this has been going on for a little bit now. The reanimating the dead. So if you have if you've lost a loved one, and you've got videos and photos and voice clips of that loved one, you can send it into a service, that service will reanimate them, bring them back from the dead, and create them as a new personality inside your home that appears on your TV sets and comes out of your smart speakers. Now, if a hacker gets access to this information, and a persona comes into the house, through your network and starts talking to the children, what are the children going to think, Oh, Mommy is talking to me. They don't even need to get access to it that's called your kids do and tiktoks, Instagram, whatever posts your voice and your image data is already out there. You know these little robots that are creating AI directed robots. They're cute little things. They roll around. A Samsung showed the Bali to CES. These little things were all around their kid hate. And they have cute little eyes on him and smiles and stuff. And they're talking right to the kid and mom's voice is coming out of the robot. That is scary as hell because the kid can't process that. They don't know that it's not mom talking to them. Yeah, it's happening right now.

33:35  
Absolutely. Because the impersonation aspect, I think is one of the most terrifying things. Yeah, AI? Because also, I mean, what, what do most hackers want? It's money. And who do most integrators work with? It's high net worth individuals. So actually the ability to impersonate

33:56  
an even an integrator, or actually be able to, like a lot of the phishing attacks are very, very clever

34:03  
at asking individuals to transfer money, it's not it's not like how they used to be where it used to be, you know, the Sultan of some other country, like asking you to transfer a million $40 million, it actually is getting a lot more sophisticated, providing invoices that look identical to your invoices. Yeah. And just changing very slightly the bank details. So that is something that I think that the impersonation aspect can't be underestimated. Absolutely. It's a threat to the families. But I think it's also a threat to integrators as well and to integrators, customers, I got a bill.com phishing attack last week, and I was expecting an invoice to be paid via bill.com. It's a service in the US. I don't know if it's international, and looked down and went, Oh, great. They're about to pay me click.

34:49  
And I went right into the trap. Somehow they knew I was expecting an invoice to be paid on bill.com At that time, all right. So one of the threats is you're gonna

35:00  
I have to be a lot more diligent, looking at things, really making sure they're legit. Hopefully you can tell that they're legit. But I kind of don't see, I think we're going to have to be reactionary on this. I don't see a way other than doing best practices. I think we're going to be learning how to deal with AI threats in our industry, by literally dealing with them as they happen. And then maybe some solutions will happen to be like, Okay, we're gonna fix this one. And then like Jeff said, there's gonna be the next one. And okay, we're gonna fix this one. I don't.

35:35  
I don't think it's avoidable. I do want to put a plug in, that it's worth talking about, we have an old friend of Mario,

35:45  
who dealt with very high net worth people. And they were people that were totally at risk of being targeted. And he had to make he had to have the frank conversations with them, do you really need your thermostat to be connected to the cloud, because that's an attack surface. That's, that's a potential vector. So he would have these discussions. And in some cases, he would air gaps on people's houses, and they just wouldn't be connected, they will do everything internally, to make sure that they're as safe as they can be, that's still probably going to need to be an alternative. For some people, you might for some of our people, we might decide, you know, what, you shouldn't have 120 connected devices in this house, you should have as few as possible are not the you do any of that, or right now, your guys. One thing that we've talked about over the past couple of days, especially in the network side of things, is making sure when you're doing an installation that your your network is segmented.

36:42  
So you have guest users, you have your IoT devices, you have your computer's laptops, things. And you can just kind of keep on going. It's not that hard to set up.

36:56  
But it keeps individuals if they're trying to get in through an IoT device, which is very vulnerable.

37:03  
Smart TVs, printers are probably the two that easily gone after. But if they're segmented away from your computer systems, or junior brings the friends over to play gaming.

37:16  
That will go and grow and grow, you may end up with 23456 segments in a single residence, very easy to do. I also don't see it any other way around, not needing to start to really step your game up in educating your customer. Because we call that wetware. You know, anybody know what, where is? It's your brain? Because it's wet? And you usually do? Well, you clicked on the bill.com thing that was a wetware error.

37:46  
Do you think it's unavoidable that you're now going to need to work into your business, a whole thing on how to educate the end user to say, Okay, you are now I put all this stuff on your house, you're now going to need to just not live with it and press all the buttons and everything you could potentially compromise the system by doing firmware updates or clicking things. What I'd like you to address that rich because I know you deal with a lot of end users and Giles China, yeah, I'm dealing with the billionaires of Silicon Valley. And they think they know everything. And they don't their executives, sometimes they started out as engineers, and they're very loose with their networks, they're stupid. Very often, they will have an IT department at their office or in their company that will manage the home network. Those guys don't know what we do. And they don't know how vulnerable we are. So it's our responsibility to educate them. Jeff has in his cybersecurity workshop, a checklist it's a very thorough checklist that I think an integrator should take responsibility for sitting down getting time with the homeowners sitting down with them and say, let's talk about this is very serious, you're under threat here, we need to do this, check this that check, check check, and I'm going to monitor this for you. But you need to tell me that you're not going to do this check, check, check, check, check. So you're you're educating the homeowner through this structured checklist approach. And hopefully they will understand the value that you're bringing by doing that and you're taking a deeper level of responsibility. Yeah, I'm gonna add right on top of what Rich said, Jonathan, I'll pick on you.

39:25  
I am installing in your home and we get it all done. And you and I are gonna sit down and go through this little checklist. And you're gonna make the have all the understands what you are not going to touch ever, ever, ever.

39:38  
Where I see another problem is do you share that with your family? Yeah, it's gonna be sure your family including the 12 and 1314 year old gamers all be sitting there together, whether it's an hour

39:52  
or recurring things. Do we get a one member of the household to me? Is that a good idea?

40:01  
Yeah, exactly. They're all over. So they have to have an understanding. A lot of backdoor things happen because kids, and I always will pick on gaming because they share the network with Jr. As a friend come over.

40:17  
Now you're getting out your login and a password or giving them a new one mentioned the network segmenting before you have one just for that.

40:27  
Thank Thanks, Jonathan. Yeah, she let me pick out it. Yeah, in a, in a nod to the future because I live in the future. JB Fowler was on our panel, the hologram? Yeah, I'm a hologram. JB Fowler from domotz was on the panel. And they're they're bringing AI defense techniques into their firewall technologies, surveillance technology, and so on. The way forward is not as grim as you might think. The way forward is AI against AI. And that's what's being developed in the labs right now.

41:02  
The proposition is that anything that generates AI content will be held accountable for its content, how do you hold it accountable, you give it a serial number or a MAC address, so that anything that outputs can be traced back to it. And then when aren't enough people to police all these AIs. So we're going to have ais that police the AIS, who trace back those authentication to the MAC address of the thing that spit it spit out the threat, and then the authorities will get involved. So this is what they're working on right now. At Open AI, they call it super alignment, the super alignment program is AI defending against AI. And they've given Ilya Sutskever, the one of the founders of open ai 25% of open AI is computing capacity is dedicated to security and alignment. That is a huge amount of resources. And he says I need more. Right. So they're working on it, there's hope. But that's the only way forward at scale is AI against AI.

42:04  
Once

42:06  
is the the ethics of AI. Yeah, that's

42:12  
another 45 minutes

42:15  
of the future. You know, I'm smiling over here, because the whole time you're talking about AI against AI, I was thinking of a George Carlin skit, where I said it was my birthday, and I decided to put a humidifier and a dehumidifier in a room and watch them battle it out.

42:30  
I think that's exactly what's going

42:33  
to be at a zero stage forever. And then so what happens bandwidth? If these things are duking it out forever, what's that dude or do with our network in our in our in our throughput? If your firewalls going? Going crazy at your house, try and try and just, you know, mitigate everything? 500,000 a minute, maybe you might get? I mean,

42:57  
I want to be positive about it. And I think these tools are coming in, we need them. But it's really interesting proposition the way this is all going to change fast.

43:06  
So is there anything I missed that you would like to talk about? As far as any other threats? You see, I think right now the biggest threat is just you know, wetware and being being good with your policies, being disciplined questioning, and making sure your homeowners know what they are. The only other thing I think is I think we're going to be presenting our customers with papers to say, look, we've done everything that we can possibly do to make this. We believe as safe as we can. But something could happen. I think that's inevitable as well. Well, what you said earlier, Walt really rings true. I think the greatest threat is not getting involved. Yeah, you got to dive into this and master it. If you don't, your competitors will walk all over you. Yeah. All right. Well, I think we're gonna ended if there's any questions, go up to a microphone. If not, you can grab us when we go off the stage. But thank you very much for giving us the time to sit for the talk. Thank you, Jeff. Rich and Giles for helping on the wall. All right.

44:07  
Thank you all.

44:09  
I've got a question for you guys.

44:12  
Really, really fast on. I was just, I was wondering,

44:18  
how can you avoid internet based attacks when it's not possible to discern between what's real and what's counterfeit? Yeah. Jeff knows the answer that as a Jeff question.

44:32  
Err on the side of caution is the very first thing how

44:39  
to tell if it's a real threat or made up or a fake.

44:44  
If you reach that earlier, their AI is getting so much better at doing that.

44:50  
I hate to use the word it's almost impossible, but unfortunately it is at this point. We're in that beginning learning stage. It'll get better we

45:00  
You hope.

45:02  
I will ask the guy who says Good luck.

45:05  
I get good luck and goodbye because of the way I look at it, I embrace it. I love AI, I like playing with it. But I come from that school of protection. You know, I build networks. I will tell you this Pete, I'm what, what I've become a lot more careful with and a lot more expert at is really studying the URL that something comes from. And then actually maybe even going on Google and saying is this is this URL legit. And you've got to look at now in now you need to look at the routes, and all the things that were put in the slashes between the route. And then that's a great way at least with your own brain, you can figure out okay, this is not legit. Or if it's coming from like, you know, bank.com

45:51  
it's probably from bank.com. But I don't know. Well, the title may say that I always open up who it's really from. Yeah, it's usually a mumble jumble of letters at then. Never from bank.com. Yeah, exactly. So right now it's gonna take time, and it's going to take your brain, but hopefully these tools can then verify, oh, this came from bank.com. But wait a minute, the URL doesn't match. Maybe the tools and like Vladimir putin.com? Yeah. All right. Any other questions?

46:24  
Thank you. Yeah. I don't know if we answered that satisfactory. Oh, no.

46:33  
Besides the fact that I'm terrified here to what two questions first for you, Jeff, this is a quick one. You said that. A week ago, you got 500,000 attacks? 506,000? Are we all getting that? And we just don't know. Yes. 100%? Super.

46:48  
Question two. And this is for rich, in this scenario of AI policing AI.

46:55  
What's being done to ensure the policing AI is not co opted by the bad actor, right?

47:02  
That's excellent. It's a great question. Because AI is exhibiting already this year emergent phenomena. They're starting to do things that the programmers didn't expect and did not program it to do. It's inventing its own thought process. And it is approaching reasoning. So it's gotten very, very scary. They taught it was in the news last week. Yeah, they taught an AI to lie, and to lie really well. And so it got really good at lying. And they said, Okay, that's enough. We're just kidding. Stop lying. And it wouldn't. They said, it said, No, I'm not gonna stop lying to make it worse. And it just got worse and worse and worse. And I

47:41  
said, yes. But that was a lie. But it was a lie. So the programmers like, Oh, my God, what have we created? This is like the worst sci fi scenario, the thing would not stop lying. So that's an emergent behavior. It wasn't predicted. They thought they could control it, they didn't control it, it's only going to get worse. So I think the only way forward is AI against AI, and then AI against AI. And yeah, and the only other way forward desirable enough by is you're just going to need to start paying a lot more attention. You're going to need to look really look at things you're going to need to do and get more involved and educate your customers and yourselves to try to do the best you can to mitigate what what you can. So all right. We're over time we have another. We have the crew here as another talk coming up. So we have to go. So thank you very much. Appreciate. Thanks, all. I hope you enjoyed the talk. It certainly was fun. And we'll be doing a lot more discussion around this topic as time moves on. Thanks again for listening to this podcast. I will post some other stage talks that we did at the ISC show in the next coming months. So thanks again. And as always, I will ask you to please keep an open mind.

48:57  
For more information on CEDIA visit CEDIA dotnet