Fighting AI Scams and Agent Impersonation with Memcyco

Hey, everyone. " Hey, my next guest is Israel Maisen. Israel is the co-founder and CEO of Memsico.

Hope I got the name right. Israel, I got it good? Yes.

Perfect. How are you? " Thank you.

Thank you for having me today. My pleasure. My pleasure, honestly.

Israel, I mentioned you're the co-founder and CEO. I always like to get our audience to understand who they're watching, who they're listening to. Tell us about your journey to co-founding and becoming the CEO over at Memsico.

So, it's the fourth startup that I'm already doing. In the past also, I was in public in the Nasdaq. So, me and my team, we working more than 30 years together in building high-tech companies, especially cyber.

So, actually, we founded this company about four and a half years ago, and after we saw a huge problem in the market, that we saw that also we actually fall to these scams, and there is no real solution, and we gave our credentials, we're losing money. So we started to research it. This is actually how we see, because it's happened all over.

Everyone almost see it. Absolutely. Now, when did you start this?

At '21, about five years ago. So this was pre-ChatGPT, pre-AI- Yeah ... kind of bursting on the scene.

But stealing credentials and this kind of problem, look, I've been in security 25 years myself, more than 25 years now. This has always been a problem, quite frankly. AI has made it a little worse, maybe.

More than a little. Yes. AI's made it worse, but it's been a problem.

So let's talk pre-AI. You see this problem. You've got a team that you've been working with, as you said, this is your fourth company, and you're going to tackle this.

You're going to tackle the problem. Yeah, so- Pre-AI, how? Yeah.

So first, AI did it much worse, not just worse, much worse, but I will talk about this later. But actually, it's happened many years, and the hackers become more and more sophisticated and do it much faster and much more, that actually what they are doing, and this is what we saw, that we lose our credentials. Suddenly someone enter to our bank account, or someone actually transfer money or buy from retail that we are buying.

Suddenly we see that there are invoices that we didn't do. So you see it all over. And this is, as you said, before even AI.

Now it's just becoming more sophisticated and much larger scale. But these scams, they started many years ago, and we saw that it's not real solutions to these scams. This is why actually we developed the products and founded Memsico.

Absolutely. Now, let's talk AI. com or what's the website, Israel?

com. Yes. That's the best place for people to go find out more.

They can get started, everything, yeah? Yes, yes, exactly. Yes.

So you say it's not just a little worse with AI, it's a lot worse with AI. Yes. What makes it a lot worse?

Because the scammer use AI even before the organizations. They don't have all of these risks and TPRM and all of these procedures that organization, they can do whatever they want. So they are creating much larger scale.

If before you saw 500 attacks or 500 fake sites or fake social media, whatever, you can see now 1,000 or 10,000. So it's much larger scale, much faster, because now that AI creating it for them, they can do it much faster. Secondly, there are AI kits, so someone that doesn't understand even what to do, use the AI kits and create this scam.

So you don't need to be just so sophisticated hacker. Everyone can do it with AI kits. So this is the main issue, that AI can create for you much larger scales of attacks and much easier for you to do it.

And this is the main issue. It is. So I call this AI scale, and there's two parts to AI scale.

One is the speed, right? Yes. They can turn these out now very, very quickly.

But not only could they do it quickly, but they could also turn out the depth of them, right? Where before they would turn out 10, now they turn out 10,000. Yes, exactly.

And that, from a defense point of view, right, it's a very different thing to try to stop 10 attacks versus stopping 10,000 attacks, and they're coming at you faster, right? Yes. And to me, that's the nugget, right, the heart of the problem is this AI scale.

Yes. And in my mind, Israel, and I'd like to hear your thoughts on it, you can only fight this with AI. In order to defend against AI, because of AI scale, you can't do it with just humans.

I think you got to use AI. I wonder what you think. Yes.

So first, even before you need to do it everything automatically, it's not that human can do it, and then manually. So this is part of the platform that we developed, that today we have six products on it, but different type of vectors of attacks that we detect. But anyway, it must be automatic.

Now, if you use AI, you can do it much better. You still can protect, but if you use AI, you can do it much better, because then it's like AI against AI. You know?

And the most important to understand which type of attacks, what they are doing, because some of them are traditional, some of them it's completely new vectors, and then you need to catch them in different way. So yes, when you use AI, it's much stronger to do. You can protect, but it's less stronger than protect with AI.

Absolutely. Absolutely. So talk to me about how Memsico is using AI to help shape digital trust.

Okay, so we are doing it in two ways. One, because there are some customers, they don't want you to use AI because they are afraid that you'll use your AI on the product. So what we did something smart.

We have a product that you don't have any AI, that if you don't want AI on it, okay. We know, as I mentioned before, we can do it without AI and with more sophisticated algorithm, regular algorithm. Now, and completely separately, if you want to use our AI capabilities, you can test it if you want for a long time, because they are more concerned about AI.

So, and after we pass all of this, then they can use, and then the AI can do as you said, it can collect many more data faster. It can correlate between data, because to protect and detect attacks, it's correlation between many events, with many sources. So if you want to do it in real time, like in milliseconds, and you use AI to correlate between all of these events, to collect them from many, many sources, and after the sources and the detection, the most important is the correlation.

Because the correlation that, oh, this for sure will be attack, or this victim for sure will be a victim, and we must protect him before anything happen. This is what we do, and this is most important, how you actually can detect before attack even started, and we can detect most of them before even attacker started. Now, AI can do it much faster and much smarter.

Absolutely. Absolutely. How's this productized, though, Israel?

So people go to Memsico, walk me through the path to- Okay ... being able to do this. Okay.

So we developed a platform. So it's a suite of product today. We started with one product in the past, that is more website impersonation protection, that you can detect all of these fake website, doesn't matter where they are, how they create them.

Then we added to this all the AI kits that we know to detect AI kits. It's very important today. A lot of these attack coming from AI kits.

Then we added another product to this suite that is many other vectors of account takeover. It doesn't need to be just from fake sites. It can be many other vectors.

And then we added also social media, because a lot of attacks starting from social media, fake social media, fake profile, that they send you fake advertisement. So we have also product that know to discover and detect all of this. And then we added, what is very important is to send the signals to the fraud engines of these organizations.

Because to send a signal, this will be an attack or these will be victims. And also another product and vector that we have that we know to discover remote access, that someone for remote access to your computer and do the actions for you. Now, actually the most important is the strength, is the database.

3 billion devices, and some of them hacker devices, and it's actually going every milliseconds. Now, sometimes we can see that this is already a hacker device attack one of our customers, so immediately when we see this device or IP or other in other customers, we can see immediately that it's a scam without any other correlation. But sometimes not, and then we need to correlate between events.

So it's a lot of signals, a few products that are on the same platform, so that we correlate. And the most important, when you install it once, that's it. You don't need to install every product again.

Just activate a new product. Got it. So I got to ask a real question for you.

I think there's two wars you're fighting here. The first war is the war you described. Right?

Having real technology solutions that'll help shape digital trust, that'll fight this front. Yeah. The second war is a harder war, and it's the public opinion.

It's what the perception of the market is, right? And the perception in the market today, we were talking off camera before we got on, the perception is people, they don't trust AI. They're afraid of it for a lot of reasons.

It's taking their job, it's disrupting this, it's going to cause that. And then they're seeing all of these scams, because not all the scams are fantastic. A lot of them are pretty crappy.

But they just, the court of public opinion is what can we do? How can we gain confidence? How can we be sure what I'm watching, or what the email I got, or what I'm seeing on my computer, how do I know it's not fake?

How do I know it's not some AI scam? Is it just, well, leave it to Memsico, we'll make sure it's not? How do you win over the confidence of your users?

Yeah. This is a very good question because it's not just what they see today. 3 billion agents will transact for the users.

So, user more and more will do actions with AI, like book tickets, flight tickets, like transfer maybe money in the future. Some of them already doing it, but a lot, as you said, very scared from this. Because, okay, well, I don't know what.

So this is exactly, and they will feel more confident in the future when they see that the organizations protect them. If you want to access to your bank, they know that this bank doing everything that they can and use all the cyber protection to protect their customers. By the way, regulators, regulation today also force the customers, the organization to protect their customers, because that it will be on their responsibility.

If not, they can get penalties, not just reimburse the customer, but also get penalty that they didn't protect the customer. So it's become very visible, and it's true, because today people are afraid to transact with AI. Yeah, there are some that doing it, but it will grow significantly because this is where the world going.

So I think this is what we are addressing also, agent impersonation, fake agent, fake agent identity. It will be critical in the next few years and it will grow like this, and then you'll use it. But I think it will take time that people, they will start feel comfortable when they believe that the organization, like their bank or airline or retail, using the right solution to protect them.

Like in the past, they was afraid to transact in websites. But then they feel more and more comfortable because they say, oh, okay, there are some protection. Still, we see a lot of scams, but there are.

They need to feel the same with AI. There are organization invest a lot and buy the right products like Memsico and other to protect us. This is, I think also where they'll transact.

They'll go to transact with AI agents and transact with them, with organizations that they feel more comfortable. So I think this will bring even more business to organizations because if I feel or the public feel, oh, this bank or this retail or this airline do everything to protect me, I don't afraid to transact with agents, this, I think will make them better, but it will take time. It will not be in months or months.

It will take some few years, I think. I agree with you. The part of the problem, Israel, is we're in uncharted territory.

No one's done this before, right? Exactly. And so we're going to have to see how it plays out.

Yeah. And the hackers much before all the customers. Yeah.

They are already prepare all the attacks, before the organization protect themselves. Absolutely. Israel, we're about out of time.

com, just the way it's spelled on the bottom third of your screen that you're seeing now. Keep up the great work, Israel. Look, as you said, it's going to be, these next couple years are going to be very interesting as we transition- Yes ...

and people's attitudes and everything matures. But in the meantime, it's companies like Memsico that are going to try to keep us safe and get us through this little transition here. Best of luck.

Thank you for coming here on Techstrong TV. Thank you so much for having me today. Thank you.

Always a pleasure. Israel Maizen, co-founder, CEO, Memsico, here on Techstrong TV. We're going to take a break.

We'll be back.