 I'm Jay Fidel. This is Think Tech Tech Talks. And we're going to talk about the cyber criminals who take advantage of people now, why and how. Okay, with our old friend Attila Serres of SIPAC. Welcome back to the show, Attila. My mind is focused on our preliminary discussion, which I'll call it ruminations and reflections. And I was talking about a company that makes software that is unpopular in the U.S., which seems to be eliminating from Russia. I don't like the idea of supporting Russia. And you were talking about chat GBT and how far it's come in terms of usefulness. Why don't you go first? All right, so we're going to talk about that. You got it, Jay. Well, GBT is fun. The best way to compare GPT to another well-known piece of technology that's changed the world is email. I mean, there was a time when companies could survive and thrive and do really well without email. Make phone calls, send to fax, whatever. And then all of a sudden one day, maybe early to mid-90s, every company had email. And if you didn't have email, you weren't competitive. What it's gotten popular for is foraging student papers or replacing coders and attorneys, right? Stuff like that, vocational or highly skilled professional jobs. But what's happened with GBT now is that it's gotten a little bit better. And if you're not using it in your company, in your business, in your creative work to solve problems, you're kind of falling behind. So six months ago, Jay, when the premier chefs of New York have taken this to heart and they've gone full out and they've been able to create wonderful creations based on ingredients in their kitchen because the GPT engine has just gotten a lot better at creating a way to create food. In the same way, the writer's strike, of course, went on for a long time. I heard it just ended, which is great news for the writers. I think that what was fundamentally at the core of the strike and what was really the contention was GPT, the fact that an executive, instead of going to a group of individuals to ask for input or ask for advice on what would be some good ideas for a show, it could ask GPT, they could farm out the initial pilot episodes to some low wage writers. And essentially the person who created the idea was an AI engine. So that would cut out that entire reciprocal process of being able to gather royalties from coming up with a good idea for a good show. So GPT is fundamentally threatening many people in their professions. However, GPT is not perfect and it still requires a human being to pilot it. And that's where, get back to the original point, which was email. So email was something that was a wonderful piece of technology, but someone still had to compose the email messages and send them back and forth and have an intelligent conversation. And GPT in much is the same way. So using it in your day-to-day work, like we use it quite a bit, can yield you several results depending on the type of work that you do. If you're involved in creative work where you have to, let's say, create emails or templates or marketing or graphics, all that can be assisted with GPT. One of the challenges we had in particular when we made, for example, SIPAC.com, which is our website, we wanted to show friendly local people engaged in business. And from a stock photography standpoint, it's pretty difficult. In fact, walking into a local business here and taking pictures, well, that's kind of invasive. So we use GPT to feed it some good prompts and tell it to create a Hawaiian-themed business with having the right people interacting in this certain way. And it was fun. We could do all kinds of stuff. We made them shaking hands and we even made them do breakdancing just for fun. And you can do all this stuff with GPT. And we had some wonderful stock photos. And you're going to see that on our website, because it reinforces the fact that we're a local company without having to go overboard with stock photos. Same thing with creating content. And this is where GPT-4 really comes in. And this has just been in the past week or two. They've linked now GPT-4 with the internet. And you're wondering, like, how does that matter? Like, what's so important about that? Well, if you ask it to query a website and say, hey, go to, let's say, thinktecawaii.com and tell me what that's all about. What does this company do? It can tell you what Think Tech Hawaii does. Furthermore, if you're designing content for that website, you could say, well, what would someone visiting thinktecawaii.com look for in a company prior to donating or viewing one of their programs? They could lay that out too. Frequently asked questions. Can you write some frequently asked questions based on this information? Yes, I can. Right. And so now you have this content creation engine that would have taken you hours. And there's probably some common sense, but could really streamline some of that process. So when we were talking before we jumped on the air, we're talking about this video editing platform. And you said I had some concerns about it being Russian ties. Well, I don't know. I could, you know, Google and read some websites or I could just feed into GPT and see what it says. And it did that very thing. It went to the website. And then it said, hey, let's look on some other being search results because it's tied into Bing. So it looked on a crunch base and a bunch of other different sites to find out who runs the company. And well, I think we just went on the air right before I got my answer. So but it's it's out there. And the best part, which is so much fun. I don't know. Jay, have you heard of the movie? Her? No. Okay, I'll give you the 10 second summary. A guy is depressed. He wants a therapist. The therapist is Scarlett Johansson's voice as an AI. And she helps him process some stuff. Of course, in the end, the AI as in all movies, decides that it's better off without people and leaves. But not before, you know, endearing him and falling, you know, he falls in love with the AI engine. So chat GBT last week released a audio engine on your phone. And you can talk to it like you would, you know, a human being. And it's pretty terrifying. It is so much fun. You know, just driving along the other day with my son, I asked it to tell me a story. And it told me this wonderful story that I, you know, beautiful story about whales and aliens and all this stuff. And then I said at the end of it, hey, make me a picture of what this would look like on the cover of a cover of a children's story. And it created it. It was amazing. Shocking. In fact, I'm happy to give that to you at the end of the show if you want. I didn't have it all queued up for you. What's the name of the app? It's the chat GBT app. And yeah, we can do some, some fun stuff with it. But yeah, this raises all kinds of questions when you connect it to the internet. And do you say that you can also talk to it and have it talk to you? Yeah, it talks back and forth. In fact, right now it's generating the, just stop the generation here. Okay. So, you know, I'll give you like an example here. We were talking about Think Tech Hawaiian. We'll see if this comes through his audio. Hi, I've heard of thinktechhawaii.com. It's a pretty cool website. Can you check it out and tell me if they're a good company? Turn up the volume a little bit. So this is where we have to, you know, cut the, cut the stream if necessary. So it's gone out. It's browsing your website. Let's see what it's all about. It's a Monday, so I think everyone's using this thing. Just demoing this earlier on today. It was taken the same thing. Thinktechhawaii appears to be a reputable organization based on the information available online. Established in 2001, it is a Hawaii based nonprofit corporation focused on raising public awareness and encouraging civic engagement through citizen journalism. They aim to contribute positively to the community by producing high quality video content for Hawaii and the global audience. While Thinktechhawaii is not BBB accredited, they have a high BBB rating of A plus. This rating, along with their 21 years in business, suggests a strong and stable presence in their field. Notably, there have been no customer complaints or reviews reported to the Better Business Bureau, indicating either a lack of issues or limited direct customer interaction that would lead to such feedback. It's also worth mentioning that they are located in Honolulu, Hawaii and have been in business for a significant period starting in 2002. This long standing operation adds to their credibility. However, without customer reviews or complaints, it's challenging to assess the direct impact or quality of their services from a consumer's perspective. If you're considering interacting with them, it might be helpful to seek out more direct feedback or testimonials, possibly from their website or social media channels. It gives me an idea, actually. But that's really more than you would get in a simple Google in the past. Anyway, there's analysis there. It asks itself a lot of questions. I'm impressed with that. But your point about the internet is really, that is disruptive because before you would ask open AI, chat GBT, you would ask it a question and it would be limited to information up to what, 2021 or something. Now it's right now. It's whatever is on the internet. Maybe that's why it takes a little longer to go checks through the resources because there's a lot of stuff on the internet. But this really changes things. Before, you couldn't get current information. You couldn't get today's news. You couldn't know what was in the times of the poster, the media in general, the articles and all the journals and all that. Now, it sounds like you can. That changes the world. That is way different than even a few months ago, even a few weeks ago. Thank you for that. This is last week. And imagine, earlier today, I was just checking, I was like, hey, what's going on in the world today? And how it might relate to me? I said, well, you're a cybersecurity professional that you're in this age group. This is what your interests are. Here are some things around the world that make sense. And here are some things you can not. And then if you don't understand, because that's the thing, Jay, I got to be honest, I'm not great when it comes to current events. I depend on my wife. I say, hey, what's going on in Israel? Sorry, I don't know what's going on. I'm not a politically minded person. And so she interprets it for me and says, oh, this is what's going on, right? And chat GPD can do the same thing if there's a complex case that's involving a very complex legal something that I don't understand. I can ask it to simplify it for me. And it's going on in the news. Tell me what's going on between this case and this person and give it to me in plain English, like if I was 15 years old. Great. That's interesting because what's going on in Israel is a good example. There are people who protest in favor of Israel, that is against the Amos massacre and women, it's using people as human shields. And then there's a lot of protests in favor of the Palestinian cause. And it's hard to do what I would call balanced reporting on that, because you have moral issues and you have time issues and you have the level of brutality issues and all that. So it strikes me that when I go to a given news source and I hear them talking about say that issue, I know within like one minute, whether they are biased or impartial, I know right away. They're just stressing this or that or the other thing because I'm ahead of them. But so you could make the same kind of judgment on chat GBT. You could ask it, you know, so what's going on in Israel? In one minute, you'll know what they're getting from their various large language model sources and all that. You'll know right away whether they're incorporating a bias from one side or the other. And then you can make your own judgment about whether chat GBT is worth the price. It's free. Well, you know, if you really want to splurge this $20 a month plan, it's probably spend more at Starbucks than you do on this. But this could, it's a very useful aid. And yeah, Jay, we can, after we're done here, happy, we can try to see what it says about what's going on with that topic and find out if it is biased. I don't know. Yeah, very interesting. Well, let's spend a few minutes on what we wanted to come here and discuss. Sorry, that was too good. Let's talk, let's talk about these guys, these criminals. And I'm guessing that from what you're saying, they're mostly on the mainland or are they on the mainland and in a way, which, which steal from people who are disadvantaged, both in food and financial assistance programs, which is in the word of the article or your article, despicable, clearly despicable. I don't mind saying that. So the question I suppose is, how widespread is that? How dangerous is that? Who's getting who's being victimized? Well, you know, this always comes back to something that happened over the summer. I don't know if you remember back in August. I know it seems like a long time ago. We're only in November. But back in August, news broke that many ATM machines and some credit card machines at local supermarkets had been compromised in that someone had placed a small device over the over the pin pad and the card reader. And they have gone ahead and they harvested the credit card swipes and the pins that were entered on the keypad and they took that information and it went to who knows where, right? And when that occurs, it means that all credit card numbers or or pin numbers, ATM numbers that were used on that device are out on the dark web. And that's, you know, EBT or ATM otherwise, right? So in my article also says, Hey, if it's EBT or ATM, it's doesn't really matter. Now, you know, it is unfortunate, but it does happen. And it's not just here. It's happened all across the country. And what the I mean, if you look at it, you know, your cell phone is essentially a supercomputer and look how small that is. Well, think about what you can do with just a small transmitter to a receiver somewhere close by and then it can, you know, take that information and send it off. And, you know, the cyber criminals took that information, those swipes and they were able to, you know, sell that info on the dark web and then use it to make purchases. And unfortunately, some of our local community members experienced that because they went to the local market to use the EBT card. And the EBT cards also use a pin, same as your ATM. And when in doubt, you know, always change your pin number, you can change it all you want with your bank, they're not going to cause a fuss. And that pin, that four digit pin that you type in afterward, that's what they were able to harvest. So they were able to make, you know, those purchases as if they were the victims who were out there supposedly, you know, trying to get food for their for themselves and their families. And it is unfortunate, but they were victimized. And it's a good reminder that when you do use your ATM card, there's limited protections in place compared to say a credit card. And so you even if you have an ATM card, you do have the option to use the Visa or MasterCard feature of that card when you're at the merchant. So next time you go to Safeway, I know they're not going to like you and not going to like me telling you this. But when you when you do go to Safeway, it assumes when you've put in an ATM card that you wish to pay by ATM. It is because they pay far less per transaction than through a credit card. But if you just hit the green button, it just processes it as a credit card transaction, you get the protections that are involved with using a MasterCard or Visa as your merchant processor. Of course, it costs them a little bit more per transaction to do that, but it gives you that consumer protection. So that's a little tech tip, maybe I shouldn't repeat. So don't don't tell anyone that, but but that that that is a way to kind of, you know, stick it to the to the markets. But in the same way, like, you know, we we do our best to try to keep this stuff from happening to us. But it's really hard when you do have a a device that's physically harvesting those credentials. And like I said, it's not just here, it's all across the country. And yeah, local criminals may have harvested that data, but it definitely went on to the dark web. And then who knows where it went from there? Is there a way, you know, technologically to stop this? That's my first question. What I mean is, is there a way like to use a fingerprint? An eye, a retinal scan, what have you? Some combination of personal characteristics to to make sure that, you know, you're properly identified. Is there a way to stop this? I mean, we have all this technology. And we're talking about, I want to talk about AI in a minute. And all this technology, why can't we stop this from happening? You know, ridiculous with the password managers and the to the to stop verification and all that has got to be an easier, better, more efficient way. What is it? There is. And security is one of those things that unfortunately does create an extra step. And, you know, one thing that I recommended, at first, when when this came out, just always tap your card when you can try not to insert it. And better yet, if you have an iPhone or the Samsung's have this, you can load your credit cards into there and using NFC and the encryption that's on board these devices, you can store and transmit and pay for things using Apple Pay and Google Pay. Those are great alternatives. That's safer. Are you saying that's safer? It is safer. Yeah, it is safer guarantees that you are who you say that you are. Also on the flip side, I'm not sure if you've been to Whole Foods lately, but they have these things where you can just wave your palm over it, read your palm print, and you can pay that way. I haven't used it yet, but I'm pretty excited. So that's some new tech that's coming out. I looked it up online and you can see folks just kind of walking out with putting their hand over the over the register over the scanner there. So there's some neat stuff that is coming out. I'm not a big fan of rental scanners. I don't like things in my eye, but I'm told that they're safe. And that is one way to also verify your identity. Biometric is another way. So you can put your finger as a biometric scanner on some of these phones to unlock the phone to authorize credit card transactions. Of course, if you have an Apple device, they can recognize your face. So you can do facial recognition. It's another way to do that two-factor authentication. But in short, Jay, it's going to take the world a long time to catch up to some of this stuff. Many websites still don't have two-factor authentication, including investment sites, if you can believe that. So some of your money is best protected with a good, strong, unique password. And that's kind of hard to do without a password manager. But like I've said before, they're pretty inexpensive. They're pretty easy to use. There's plenty of good... And sometimes they get hacked, though, I tell them. They do. But to a degree, if you have good cyber hygiene in the first place, then even if they do get the data, they can't do anything with it. That's been the case with LastPass. So LastPass did have their vaults, the encrypted vaults extracted. But if you had a weak password on there, like password one, two, three, well, they're going to get in anyway, right? And then they have access to all your passwords. So have a good, strong, unique password for that. And then you have strong, unique passwords for every other website and have it manage it for you. It's pretty much... Okay, I want to put... I want you to put your AI creative hat on for a moment, okay? So here we have this problem you described. It's what you call it, the identification code that you have to punch in in various ways that these criminals can steal money from even disadvantaged people, especially disadvantaged people who may not be that computer akamai and don't see it coming. But AI is smarter than... Well, we don't know how smart it is yet. We know it's very smart. And it can be made smarter and it will be made smarter. Why can't AI be used, A, to prevent this from happening, and B, to catch those buggers, to catch them and make a case out of it and put them in prison, federal prison. I prefer federal prison. The food is better, yeah. So, Mary, why can't AI be brought out on this? Is there somebody working on this? Do you think there is? Do you think there will be somebody working on these really basic problems, electronic problems that we can't seem to shake? Wouldn't it be wonderful if AI could identify the guy who's doing it, a pattern, you know? AI is all about patterns. Talk to me. Tell me your creative thoughts. Well, I'm going to tell you a secret, Che. We've been using AI a long time here at SIPAC. And not the chat GBT AI that you're thinking about. We use it as gateway level protection, workstation level protection. So that means that it's using AI to look for anomalous patterns, patterns that could indicate compromise on the individual computers and workstations as well as at the gateway level. So anyone trying to break in, anyone trying to break out of the network, if any of that looks suspicious, that matches any sort of patterns that it may have learned about in the past, that kind of machine learning has been around for a long time. So this task-specific AI has been around decades, right? Not new. But what's new about chat GBT is that it's now also incorporating behavioral characteristics. So if we want to talk about how you can keep this from happening to yourself, then chat GBT is going to give you some good guidelines for behavior. So when I mentioned earlier about using your PIN, chat GBT can recommend to you, as I recommend as well, that when you type in your PIN, put your hand over your other hand so that it can't be seen. We're kind of doing this here so it ends up on camera. Here we go. So it can't be seen. It can recommend that. Oh, look at that. What else could it recommend? It could recommend that you change your PIN if there's been any sort of suspicious activity on your account. Do you review your statements once a month? Maybe check it out. See if there's been anything unusual on there. One of the other best pieces of advice that I've heard from other companies is that they have different bank accounts for different purposes, or different credit cards for different purposes, kind of like a football team, right? You don't have everyone playing the same position. They pay different positions for different purposes. And in the same way, you do the same thing with your finances. Not all of it needs to be in a giant pile of money. You can break it up for different tasks. That also helps to kind of segment your responsibility. Now, we do this on a network level, if you can believe that. Many folks come to us and they say, look, we've had this network. It's been here for a long time. It's starting to get a little bit old. And we understand that it's not as secure as it should be. And the kind of expertise we need and the disruption involved is not something that we want to do on our own. Can you come help us? And at that point, we'll break up the network. We'll make it so that even if there is an infection, let's say in your production environment, it doesn't spread over to accounting, right, in that kind of way. So we're talking about a behavioral change to a technology kind of problem, right? It's like watertight doors on a ship or a submarine. Exactly. Exactly. You do want to air gap different parts of your network. Want to air gap different parts of your money. Might want to make sure that your phone's backed up. I mean, I can't tell you how many times we talk to folks that say, oh, yeah, I broke my phone. I lost my phone. I didn't have a backup. I don't know. I clouded like a dollar a month. We're not talking big bucks here. What I was thinking of is the pattern that is established. It's like back in one of those crime movies, you get an M.O., right? And the police identify the M. This is old fashioned stuff. They identify the M.O. And they look in another city or state and find the same M.O. exists. And that helps them point to the perpetrator, you know? So for example, using AI or whatever system that you can develop an M.O. on a certain kind of break-in, a certain kind of criminal activity on the web. And you have a global database of M.O.s. And you find that it's coming from here. It's got these numbers in the header. And it just looks like the one that came out of Bulgaria. It looks just like that one. And then you watch Bulgaria. And then bingo, you find that somebody in Bulgaria, I like to make a movie like this, be an interesting movie. Somebody in Bulgaria is doing that M.O. So now you're focusing maybe you'll even find the guy. Maybe you'll find him by way of the M.O. despite the geography. Can AI do that? Will AI do that? Because we'd like to see some prosecutions. We'd like to see some guys tossed in jail. And that's not happening. Yeah, we did write up on this also on our deep watch blog about Nigeria specifically. And Homeland Security specifically does monitor IP address ranges. And then they do make announcements out to the private sector network defenders, such as ourselves. They say, Hey, look, these are the IP addresses. If you do see this in your network, it's bad. So make sure you put that on a block list. And and yes, AI is definitely used for trend analysis. She said, you know, take large blocks of data and look for patterns in ways that you and I can't do very easily. So yeah, it's been in use for a long time. That's as part of network defense. If you're not using AI, you're behind the times. So you know, the problem the problem is ultimately bring trotting out the report of the AI and using that as exhibit a in a prosecution in a criminal trial and indictment and trial and getting hold of this guy. That's that's number one. You know, coming up with admissible evidence from what AI can do on identifying patterns and the other thing and we know whether a trial or a fact would accept that, you know, you'll agree that technology in the criminal field, you got to get the system to agree that it's admissible. Sometimes the system is under fingerprints, for example, were known in the 19th century, but they weren't really treated as a police device admissible evidence and until well into the 20th century. And so the same thing with, you know, evidence that comes out of AI. So that, you know, and that of course, you have jurisdictional problems around, you know, global jurisdiction of cyber crime. You can tell me, but I don't think there's any organization where various nation states have signed up to participate in the prosecution of cyber crimes that are emanate from somewhere else. Do you know? Well, all I can tell you is this, when you step inside of an airplane, there is a pilot there. And we all know that the airplanes are capable of taking off and flying and landing on their own without a pilot. But good luck trying to get, you know, grandma onto an airplane without a pilot on there or most people. It's very uncomfortable. In the same way in law, AI and what it can find and help you on earth is a research assistant. It's not designed to make the case for you. It's designed to aid and assist and, you know, find the worthy bits of information that's that's buried in the stack somewhere so that it can help your argument. Someone still needs to make the argument. And as far as our prosecution goes, I can tell you right now that, you know, AI is used quite a bit by the legal community to find information, get advice, mind data, do open source intelligence. Yes, absolutely. It is being used. But there's always an attorney at the other end who's presenting the information and making sure that it's curated and it is correct and legal. We don't want to have problems because AI is fighting cases for us. I think we're pretty far from that. I'm thinking of that movie with Tom Cruise, which examined the notion of pre-crime. Yeah, yeah. What was that one? Yeah, I forgot. I know what you're talking about, where they're doing a lot of things with their hands as the interface. Yeah, yeah, that was part of it. But the part that it was, you know, memorable is the minority report, yeah. It was memorable was the notion that this machine was actually a woman in a swimming pool. This machine could identify, she was really smart, that she could identify criminal patterns in individuals and then the state would know this guy was likely to commit a crime. And if so, the state would take some action. I mean, it was really very undemocratic and it didn't have a whole lot of civil rights to it. But the study was that you can use technology, at least in the movie, to identify a crime or a criminal before the criminal has committed the crime. And I hate to even think this, but AI could do that job, couldn't it? Well, I suppose it could. I mean, it's certainly not a woman in a swimming pool. But, you know, the same way, I think we overestimate the simplicity of the human being, right? Like we have very complex creatures, we're very situational. And someone who could create a crime in one situation might not create, might not do that same crime in a different circumstance. So yeah, I mean AI could be used as a preventative measure, but I think more practically we can use it to improve our day-to-day lives, right? Like that's where it comes in. We can do so much more so much faster. Think about vaccines, think about research, think about space exploration. I mean, one of the big challenges we have with astronomical amounts of data is that it's literally astronomical amounts of data. We're talking hundreds of terabytes of data coming in that's being stored on star patterns and stuff. I mean, think about planetary defense, right? AI can find these things out if, you know, what if an asteroid is heading towards Earth? Good luck, right? But AI may pick that up faster than we could. Think about climate change. How can we analyze different data patterns and find out predictively if there's a storm coming and what we need to do due to prepare and save lives. I mean, this is where AI is really exciting. Just think about how many lives you can save. Think about how you can improve the quality of life for so many. Think about how you can, you know, work out, solve these fundamental human problems we have about food and energy shortages and how we can bring a quality to the world. I mean, that's the promise of AI. That's the excitement of AI. I really hope it does come to pass and it doesn't end up being a terminator situation, but, you know, I'm hopeful. Well, your news about the AI chat GBT incorporating the internet and being, I guess, it's the same thing is a remarkable step forward on that. Instead of having, you know, the limitation of old news to draw from or a language model that's like static. Now, now you have information that is really infinite. It's the whole human experience every moment. And no, no wonder it takes a little while for chat GBT to actually come up with a plan. But I think this opens new doors that tell us we're going to see remarkable changes. It's going to boggle the mind and the imagination to see what happens. Yeah, it's coming. We have GPT-5 coming out in the next couple of months. I mean, everything is moving rapidly. I mean, the computers that you see today are going to be totally different, even a year from now. Okay, well, I guess we covered both ends of that. We covered ruminations at the outset and throughout because it is a strain that is always present now going forward. And we did cover the monsters who would steal money from disadvantaged consumers. But I'm sure there'll be more on both ends of that. So let's plan to circle back in a few weeks. Okay. Sounds good. Thank you, Attila. Enjoy your Thanksgiving.