#253 Ivan Shkvarun: Inside the Fight Against AI-Driven Cybercrime

0:00I think that it will be interesting to know that there are 1 billion websites now and this is only 4 % of internet That's why when I even now call the internet, honestly, I mean all digital space and 96 % of the data in internet are in darknet. So we have a dark web monitoring and dark web solution and social links. And that's how we started. And that's how I found a lot of insights and that we launched in January. I called it dark side AI initiative. Why also dark side AI? It's also the story. Two years ago, we started to work with company and we even invested in this company. The name of this company, Brightside AI.

0:45So what guys are doing, they are trying to use AI to defense the employees of the companies from the AI solutions that criminals are using. Our sponsor for this episode is Extreme Networks. the company radically improving customer experiences with AI-powered automation for networking. Xtreme is driving the convergence of AI networking and security to transform the way businesses connect and protect their networks to deliver faster performance, stronger security, and a seamless user experience. Visit extremenetworks.com to learn more. That's extremenetworks, all run together, dot com to learn more.

1:38Hi, my name is Ivan. I am CEO and founder of Social Links Company. And at the same time, I am author and founder of Darkside AI Initiative. and let's introduce social links because that how it all started and we learned this company in 2015 and we were one of the first commercial company in open source intelligence area what we are doing we are gathering public available information about people and companies and helping governments and businesses to secure

2:22to secure and to make our life and businesses safety and that's how we saw that criminals started to use ai much more and early than businesses even started to use all these benefits from ai and that was one of the first thoughts about okay we should do something and i am an entrepreneur guy and that's why also the idea was okay what could we do and that's how we launched this drugs idea initiative we just launched it in january so few months ago and um what is that it's initiative that's mean that it's not a product and solution it's like non-profit project we are from one side we are doing a lot of research and we will public them for free to increase awareness about this risk we will make some educational things of course we will try to develop solutions and we're already doing that from another side the main idea to join community about that to discuss maybe to launch any standards and also to talk about that because if you will not talk about that and we will think that oh it's so scary we could not talk about that nothing will change yeah i wanted to talk a little bit about social link so you gather public publicly available data not only on the visible web, but from the dark web as well.

4:00And then police forces or governments or anybody in the enforcement space can use social links to understand criminal networks or identify individuals involved in those networks. Can you tell us how that works? Yeah. I think that it will be interesting to know that there are 1 billion websites now. And this is only 4 % of internet. That's why when I even now call the internet, honestly, I mean all digital space. and 96 of the data in internet are in darknet so that's why another name of internet surface web yeah and that's why darknet dark dark web so that's why how we started and we realized that there are a lot of different places where are the information and we should look at that i will even rephrase it like you know that we are more powerful than google engine because google indexed only data that you could see or sites that you could see with standard browsers we could find the data that are invisible for standards browser so that's the easy way to explain but the idea yes how it started that a lot of people started to put their information anywhere in the digital sphere so facebook is just the first example because we all know that this was one of the first and famous even for now social network yeah but it's not about that the story is that it was the first steps of data migration from any kind of internal sources to open sources and this also that we saw and i see even now that this digital hygiene of the people hygiene yeah it's near the zero we even don't think about what are we doing could someone see it and so on so on that's why it's usually like black and white yin and yang principle because from one side we could find that and help to law enforcement to catch the criminals because there are a lot of digital footprints how we call it and there are a lot of interesting stories how digital identificators of our of us are connected between each other like the any ap addresses emails ids and so on so on yeah so we could see the full picture of that and from one side we see it and we are trying to indicate the people also to show guys like the standard principle on cyber security don't use the same email everywhere like a login yeah because all these solutions and applications they helped for us they helped us with the one crucial thing that now login is email so everywhere your email but that's mean that it's so easy to to find you and to understand you.

7:53And we educate it. From another side, yes, we use it to help law enforcement and governments and businesses, corporate security, information security departments from the biggest S &P 500 companies to secure their businesses, to find the criminals. And also that's why we know how criminals are using that. And that's why we could help, you know, like the virus and antivirus solutions. yeah you should understand how the viruses works to secure it and that's how we help the companies to work with that so yeah we get information from more than 500 sources different sources they are all public available information so we are fully compliant here with any rules and principles here because this public available formations you could find it manually and everyone could find it but we did a automated solution for that yeah that could help you for one minute to see it if you want to do it manually you will spend days or weeks but you also could find it so that's i mean public development information not this case how do you search the dark web how do you connect the dark web to to say uh facebook uh so that you can identify people because the dark web is not indexed.

9:22So unless you know, uh, the, uh, the address for a particular, uh, site on the dark net, uh, you're never going to find it. I mean, there, there are all kinds of lists and, and things like that, but, but it's constantly changing and, and they're not very useful. So how do you crawl the dark net, identify sites and identify and then link them back to the visible net? I will try to explain you, but I will not tell all details because that's why we are the leader in OSINT sphere. For example, you have any media post or media article and you post the email. Like, guys, if you want to write me, this is my email.

10:18So it's public. We see it email. And of course, it's very easy to identify that if you are the author of this article and you posted your own email to connect with you that this is your email and so on. Just let's put it in our mind. Then let's go to dark web. The most interesting thing that in dark web, all messages are secured with special open and close key. And the idea that if you want to produce your own key, you should register the email. So you understand that even in dark web, you should use an email to register your account, to register your own key secret key and so on so on so on and of course the most professional guys they have uh one unique email for any transaction and so on so on but we are a human being and the idea that someone lazy and someone did mistake and all most famous criminals who were found, they did any kind of mistake.

11:39And that's how we could find them. So the idea that how you could analyze that this is the same person, for example, in internet, in surface web and in dark web, if you could connect that the email is the same. so if you will find the email here in the mail here you could predict that maybe the same person and then one more hypothesis how to look at that so that's why i mentioned that sometimes you you know that even i have a joke like if you are for example walking around on the street and you see any boxes for example ikea delivery box or something like that you could look at that and in most cases what you will see here the first name the last name the phone the addresses you know like dhl delivery list who are that so just imagine the personal data are just like a garbage on the street and you could just look at that and it's reality i checked a lot of times so that's how how even our solution also works.

12:51So we look at a lot of different places and combine the full picture by any pieces and each of piece, it could be fully unuseful. But when you look at that altogether, it's like, wow, yeah, inside. Yeah. Yeah. And how do you crawl? Do you have crawlers that are going through the dark web autonomously? Or is this gathering this data? Is it a manual exercise where you have an army of people searching through the dark web? No, it's all automated. And I will tell that there are really different methods and types how we are doing that. In like 90 % of cases, we are doing that by our own and this is our own solution.

13:57and of course there are some companies on the market who are also doing that and we are using and so we buy the services and there are a few really famous companies who are doing that we have more than 500 customers more than 90 countries and we decided to and we are still doing that because we did the first project and we decided to check how many uh breaches in darknet connected to corporate emails and we found a lot even from ministry of media ministry of defense ministry of interiors the full country what does mean that the devices of the employees of ministry of defense some countries are infected and every day they have breaches just imagine what is going on yeah of course we notified them we tell them they did some things with that and that's why we just continue to do it monthly and we decided that we will make the deep research and i will also uh use it in our dark side the initiative to show that guys we could find it so just imagine what criminals with ei could do with that and why you are not doing something really crucial with that yeah build the future of multi-agent software with agency that's a g n t c y the agency is an open source collective building the internet of agents.

15:42It's a collaborative layer where agents can discover, connect, and work across frameworks. For developers, this means standardized agent discovery tools, seamless protocols for interagent communications. For developers, this means standardized agent discovery tools, seamless protocols for interagent communication, and modular components to compose and scale multi-agent workflows. Join Crew AI, Langchain, Llama Index, Browserbase, Cisco, and dozens more. The agency is dropping code, specs, and services. No strings attached. Build with other engineers who care about high-quality multi-agent software.

16:34Visit agency.org and add your support. That's a-g-n-t-c-y dot o-r-g agency.org and support the project yeah and just for listeners you mentioned osint a little while ago that's open source intelligence uh and that's that's what social links is doing you're relying on uh open source you know social media platforms uh news articles academic papers government reports and the dark web web content uh to to to create a visualization of how everything is interconnected is is it uh do you do visualizations or is this uh just stored in a relational database or something uh i should mention that we don't collect all this data it's illegal oh so we are our solution work works in real time so that's why i mentioned like we are like a google i mean that it's changing in real time so that's also we are trying not trying yeah of course we are fully compliant and this is one of the main thing that we did not store it and collect so that's why this infrastructure is even so unique yeah that it's in real time in seconds sometimes in maybe minutes if it's really huge query and so on that's why we are doing all steps of intelligence cycle we are gathering we are also using our own ai models to speed up investigation to analyze it all texts speech analysis visual recognition and so on so on and then we have different visualization tool like maps tables graphs uh timelines yeah also to make an analytical report easy so that's one of solution and i want to mention that 10 years ago we started with the main idea that all seemed uh and this is like the the market the most solution i know but But now more and more I see and think that all things, it's mostly methodology, like a mindset, you know, the steps that you should do to have some values.

19:29And now I realized, and we also did a few steps for that, how this methodology could bring the value to new industries. for example not only to securities yeah for example we launched spin-off company for financial sector so we understood that in emerging markets there are more than 80 percent of the people in countries they could not use standard digital banking because they don't have any credit score yeah nothing so they really could not use these benefits and we understood how to make a credit score based on alternative data and the banks and loans platforms and so on they found the benefits here and also there are a lot of ideas that we are doing with digital risk protection in healthcare education transportation and so on there are huge research by mackenzie that showed that open data could unlock from five to six trillions of dollars and they called it like trillions of values to these new industries if we find the way how to use it and we are doing that step by step so that's why I think that this is methodology the methodology the mindset how to use data because you know also the interesting fact that we found that more than 50 of companies they don't know and they didn't have the benefit from their own data so just imagine yeah they are producing a lot of data and they are not using their own data for any decision making steps So not external data, they are not using their own data for decision-making.

21:43So it's a huge problem with all of this data. So when I try to explain a lot of companies that guys, there are a lot of interesting facts, insights, and like in cybersecurity, you know, the guide Gartner mentioned that 10 years ago, the most amount of risks was inside the land IT landscape. but nowadays even more risk outside the they're outside the it landscape so that's mean that all the social engineering staff approaches and criminals they are trying to collect information about the victims and then use it so that's outside and all the cyber security security operation center and cm solution they should start to work with public development information They're not doing that for still now.

22:38So a lot of things to be done to bring, to educate, to explain that, guys, this data could help you. And this also, we will try to do it. Build the future of multi-agent software with agency. That's A-G-N-T-C-Y. The agency is an open source collective building the internet of agents. It's a collaborative layer where agents can discover, connect, and work across frameworks. For developers, this means standardized agent discovery tools, seamless protocols for interagent communication, and modular components to compose and scale multi-agent workflows. Join Crew AI, Langchain, Llama Index, Browserbase, Cisco, and dozens more.

23:41The agency is dropping code, specs, and services. No strings attached. Build with other engineers who care about high-quality multi-agent software. Visit agency.org and add your support. That's A-G-N-T-C-Y dot O-R-G. You're based in Amsterdam. I think you studied in Moscow and then Germany. Is that right?

24:19How, I'm sorry, I'll cut this out. but uh social links social links is a leading software provider of uh for osin for open source intelligence are you providing uh is it a sas platform or is this a software package that that people buy and install and use we have different options and basic options is our solutions so that's how our we are working only with b2g and b2b companies so we are not working with individuals also why because it's so difficult to to check them yeah and I won't be sure who are using our solution of course because it's really powerful yeah and we have strict kyc and all these uh things that should be done because our holding company in us we have investors and we are one of my dream uh to be presented uh at nasdaq yeah so we are looking at it super seriously and that's why we have a SaaS solution, of course, but for customers who want to have their own on-premise solution, we have it.

25:47So we have self-hosted version. It's something like the hybrid between SaaS and on-premise. And we have fully on-premise solution. In most cases, this is national security projects. So we started with fully SaaS, but then we saw the requests from customers that want to have another type of these implementations, and we did it. So let's talk about the Dark Web AI initiative. So you have this solution to help companies, government agencies, law enforcement agencies explore links between potential criminal organizations and individuals that you can identify on the visible web. uh tell tell me again about the the what the dark web initiative is who's involved in it and uh what you hope to achieve by it yep uh just let us specify so we have a dark web monitoring and dark web solution and social links and that's how we started and that's how I found a lot of insights and that we launched in January I called it dark side dark side AI initiative why also dark side AI it's also the story that two years ago we started to work with company and we even invested in this company the name of this company bright side AI so what guys are doing they are trying to use ei to defense the employees of the companies from the ai solutions that are using uh that criminals are using and that's why they called their company bright side and this was so cool and then you know when i spent uh in 2024 a lot of times in us yes i am basically in amsterdam our company in us and i was surprised that most of people and companies are talking about how ai will help them to decrease the costs or to automate the processes or blah blah blah so you know you see it and you know it all crazy things about ai Everyone is talking about AI.

28:2810 companies in YC Combinators are doing something with AI and so on and all. Yeah. But one more time, just no one are talking about how AI, how criminals are using AI. And just the fact that the cost of one fraud transaction now less than 10 cents. so in five years ago four years ago they were needed the criminals 100 guys in call center who were calling and sending you messages now they should have two developers and few AI agents that's all you see and they could attack thousands of peoples in just one hour so that's how AI I changed the criminal sphere also, not only that we see. And that's what the last point when I realized, okay, we should do something and we should name it.

29:27And we decided that we will have few streams here. First of all, of course, we are doing our technical solutions. So it's not the initiative. It's not about to talk or to discuss. We are really developing the solutions. and the solutions in the sphere of modern digital risk protection. So we are helping companies to monitor the digital risks about their digital assets. We were thinking how to increase awareness, how to show the people about their threats. And we decided to do a lot of research. And we will public them for free with a lot of statistics. And we are doing that by our own. From another side, we are doing that with partners.

30:13and for the i don't know this few two months i received more than 50 requests on our site to collaborate together to do something and that's i think really cool so no special advertisements you know like nothing to do but people decided and voted that we want to be a part of that and we are talking with them what to do so we are open to do any research together to educate the people and community the one more thing that we decided to do to think about real solution that could help to to minimize this misuse of AI for example any standards or anything like to label AI AI, for example, and some methodology and solution how the AI generated content should be labeled.

31:16Yeah, just because it's so really difficult to understand. So some real steps that we could do with the businesses and governments to learn some standards. I don't like the word rules you know because i think in this sphere rules are not working because what that means rule so someone decide that we could do that not we i mean that uh that could be done that not could be done and why do we think that it's right now in so modern technology world yeah So that's why I think there are no rules and there are nothing to do with like to stop the AI development. You know, sometimes we hear it on the market.

32:05No, let's stop AI development. We could not stop it because it's a curiosity. A lot of people doing that, like, okay, what could be done more? Or it's so interesting. Yeah. So I mean, some standards, some principles. Yeah. so that's few steps that we are trying to do and of course we are trying to do that with partnerships from companies and organizations to do but just two months passed yeah but even uh i see these results even now that uh the the interest are huge and we will try to keep it forward. And how do people get involved in the Dark Side AI initiative? And I'm sorry, I called it Dark Web AI initiative.

32:56But how do people get involved in the Dark Side AI initiative? Is there a website? Do you have conferences? Do you have webinars? I mean, how are you communicating with people and brainstorming? Yeah, we started, of course, with the website DarkSide.io and LinkedIn page for this company. And we started with the interesting content to show some numbers, to show, for example, the cases that more than 90 % of people could not understand where AI generated content and we are not even like I think more than 99 % could not just realize yeah and of course we could not do it by alone so main idea to engage involve community and I think the main thing from my side try to lead it to engage people and help them also to do something yeah and and the idea then is you you you build this community there's exchange of ideas within the community and then will you have uh you know maybe uh committees within the community working on specific solutions that are for for problems that are identified i mean how do you see it developing yeah there are also a few uh streams that we have uh we are developing our solution uh this modern digital works protection and we are ready to invite experts who are interested in that to develop with us or to test it so we already did few projects for free for big companies to show them the benefits from one side from another side of course to test in reality could we bring the value or not the cool result is that the feedback of customers that we are outperformed then some competitors on the market so we just spent uh half a year for team for people to develop that and we already outperformed.

35:37Why? Because this awesome methodology and expertise for the last 10 years because of that. From another side, I already started few projects with individuals and companies from different markets, even in Japan, who have their own idea to do something. And we will help them. We will help them with technology. We will help them with our expertise. sometimes we are even ready to invest our money in that if it is really interesting and so on yeah you have some data about the increase of ai driven fraud i i was this your number to expect it to grow uh from 12 billion in 2023 to 40 billion of uh fraud by 2027 is that is that your number yeah and it's growing it's growing extremely because even you know like the most the biggest fraud in finance when ce4 paid because he received even the call from ceo zoom call from ceo yeah it was messages and then it was called and he did not realize that it was AI generated CEO on the call by Zoom and so on.

37:15So just one more time, imagine what's going on. Yeah. And yes, I see that it's growing. And of course, from one side, the most, the areas that are most dangerous, of course, it's finance, because it's a lot of, it's about money. So of course, it's one of the targets for criminals. From another side, it's all that um i about content so that's why media and all these uh public institutions but from another side we see that it's not only about finance because it's all industries are connected to that and i think that criteria even not about from criminals are not about like banks they have money we should do something yeah they are looking at enterprises and smaller media companies that have more breaches that are not secured and then choosing the goal for them and doing that so unfortunately it's not like something that only the most richest and biggest company could be the targets no that we see uh and why also you know when you when criminals had the hundreds of people to did something the ROI of that should be bigger.

39:01Yeah. Now, few developers in the UI decrease the costs for any attacks. And that's why they could choose not only enterprises for that. And that's the change. Yeah. Yeah. So you're inviting enterprises, researchers, policymakers, individual experts, etc. to join the initiative. And to do so, they can start by visiting www.darksideai.io. Is that right? yes really we'll be glad we have a lot of experience and expertise in all this data so if you have ideas or if you see any risks and you want to research it we will be glad to help you and to do something together to benefit and to secure the world the digital sphere and our lives definitely yeah and then social links uh how does somebody contact social links if they want to use your solutions the same thing that social links dot io the website and you could also send us the request and you could find a lot of open cases materials about social links so honestly it's much much easier to find us and to see we produced a lot of content landscape documents so there are a lot of things how you could find us for the last 10 years and that's why the dark side we just launched two months that we are on the market yeah and this is this is a global initiative dark I'm sorry, Darkside AI.

41:03Is that right? You said that Social Links has customers in 90-some countries. How many countries are there participating in Darkside AI so far? Is it primarily a U.S. initiative? No, there are no limits. And that's also one of our main expertise, that we are working with more than 90 countries. And we see these sometimes some specific cases, but sometimes not. so why also because you know uh i was asked interesting question a few weeks ago who are that criminals so you know like who are they what mean criminals and my one of the person third that it's like i don't know like few students who are doing some bad things but criminals it's the same huge enterprises with the legal department sales department marketing department security department in reality it's a huge organization with a lot of professionals and if you will look at their structure in this company it's the same that you could see in the biggest s p 500 companies they're super professional guys and based on that when they are doing something in most cases they are also global they don't want to look like at one country in one market why yeah this is digital sphere you could do anything from anywhere so there are a lot of of course there are a lot of uh regional specific projects uh and one of the last specific that i saw it was connected to japan yeah because uh there are a lot of specific in japan in reality even for example language yeah and so on so on but from another side even uh japanese organizations are not secured now so i mean that there in i i could not tell any limits uh or ideas that someone are safety in red i don't see it so that's why it's fully global there are no any limitations and i see that we could bring that we already did it in so many countries and that's why that we could bring value to any new.

43:56So how many, there are something like 250 countries here in the world. So, okay, we already helped 100. So just 100 more. Okay. Okay, well, let's leave it there. Is there anything I didn't ask that you would like to say to listeners? uh i think that one thing just to finalize it and it's my uh one of my favorite and that i believe more that we should educate ourselves yeah so from one side we or a lot of companies are trying to educate you to explain you about the threats about cyber security basic things about digital hygiene and so on so on but we should realize it for example i have two children my daughter is eight and my son 13 and i already helped my son to find the cyber security courses for children to look at that because they will live in this sphere in this world and we should start it for example the same like in my in my children's international school they watched the cinema there was a cinema on netflix about social media platform how they influence on people how they influence on children because know there are a lot of children who are fully disappointed because no one like them what's mean no one like you because i have not so many likes uh in any social media so just the mindset of children yeah that they are not successful because they have a lot of likes they have a small amount of likes here so that's and it's cool example also that if you will start with your children to look at this fear, you will find a lot of interesting things for you.

46:12So that's why you should think about yourself by your own. And that's one of my advice started immediately. Now don't wait anything. Our sponsor for this episode is extreme networks, the company radically improving customer experiences with AI-powered automation for networking. Xtreme is driving the convergence of AI networking and security to transform the way businesses connect and protect their networks to deliver faster performance, stronger security, and a seamless user experience. Visit extremenetworks.com to learn more. That's extremenetworks, all run together, dot com to learn more.