AI’s Role in Data Security and the AI Regulatory Environment

00:08Welcome to NASAC Trae Talks,

00:11where we meet with the top thought leaders and strategists in TradFi Digital Assets,

00:14technology, and Financial Planning.

00:16I'm your host Jill Malingerino,

00:17and joining me on the desk at the NASDAQ Market Site is Thomas Jewell,

00:20CEO of IPA, Carrie Collegia,

00:22CEO Versap and Ronan Cohen,

00:24vice president of Strategy duality.

00:26We're here to discuss AI's role in data security and

00:29Data Privacy and the AI regulatory environment.

00:32It is great to have everyone with us covering my favorite subject on Trade Tox.

00:36Let's go around the horn quickly, Thomas. We'll start with you.

00:38Explain to us where IP fits within the Data Privacy and AI ecosystem.

00:42Yeah, thanks, ill. Thanks for having us.

00:44So InPay is a cross border payments company.

00:47We do cross border payments all over the world.

00:49So we connect different different senders and receivers.

00:52We try to do that safe,

00:54fast and secure, which I think is part of what we are here to talk about today.

00:58And for us, it's about technology,

01:00it's about network, and it's about compliance.

01:02And we built that for 16 years,

01:03and we believe that we are the leaders in that technology.

01:05All right. And are versip?

01:06Sure. So ersaps in the business of offering accounts receivable efficiency suites.

01:12Ultimately, what that means is we try to make it easier for

01:14businesses to get paid for the goods and services that they provide.

01:17And we do that through the process of automating invoicing,

01:21moving b2b payments, and as well as reconciling cash invoices, using AI itself.

01:26So you can imagine with every touch point when money moves,

01:29data moves, and therefore,

01:31security compliance and regulation is critical.

01:33Alright. And duality. Thanks for having me.

01:36Duality is a secure data and AI collaboration platform.

01:39We're used in the public and private sector by organizations that want to

01:43jointly analyze very sensitive data and derive outcomes and insights with each other,

01:47and it's used by collaborating partners and sometimes even competitors to do that.

01:51Yeah. Well, it's so interesting Rod in because you hear competitors,

01:54consumers, you know, your own tech stack and so forth.

01:56And then AI, it's used for really,

01:58really good things, and it's also employed by bad actors.

02:01So what is its role in data security?

02:04It plays a huge role, and data security plays a huge role in AI as an enler to AI.

02:09You know, privacy and security are really the main blockers in terms of leveraging data.

02:13I think what you'll see is that AI application providers,

02:17the way that they're going to have to differentiate themselves is how they

02:20can access sensitive data to make their models better,

02:23to work more closely with their customers.

02:25The big challenges there are security and privacy.

02:27Right. And especially, you know,

02:28when you're thinking about cross border payments and so forth.

02:31And from what I understand you operate in virtually any place

02:33around the planet where you're able to.

02:35Yeah. That's correct. Yeah. So I can go totally back with Ron is saying he,

02:39because it's all about how do you as an industry?

02:42This is not about in pay and not about specific companies.

02:44How do you as an industry build a safe environment in

02:46which the good guys can transact and when the bad guys are kept out?

02:49And I think what you're saying that's exactly what we need to do we need

02:52to think about it as how do we competitors, partners,

02:54whatever we call ourselves,

02:56actually act together to secure that

02:58the payments you deliver are actually done to the right people.

03:01Yeah, I would add to that.

03:02I think, you know, we've really shifted in our society and in the commercial ecosystem,

03:07too, it's not about how do I win?

03:10It's about how do we win,

03:11and the world is moving so fast and technology is moving so fast,

03:14like, you can't keep up Jail anymore with all aspects of your business.

03:17And so if I can work with a competitor or a partner

03:20to share data signals that are done anonymously with

03:23keeping the consumer privacy regulation in mind to help

03:27service them better and to help protect against bad behaviors and efarious activities.

03:32Like, I think that's something we would all want to opt into.

03:34And but finding those right platforms are

03:36difficult to And that's essentially what duality does.

03:39You sit in the middle of the data custodians, the compliance teams,

03:42and then being able to build these applications leveraging AI.

03:45Exactly. Absolutely. And you know, there's a dilemma,

03:48and there's always a trade off because it's a we versus me like aryan Thomas brought up.

03:53But at the same time, you still have to protect yourself,

03:55you have your IP, you have your sensitive data,

03:57you have your proprietary data.

03:58How can you balance that?

04:00And that's where privacy enhancing technologies come in and secure collaboration in AI

04:05comes in to be an enler

04:07to these outcomes and to these insights and to keep in your world,

04:10bad guys out and the good guys in.

04:12But how do you do that, like, with who owns the IP of the model that's created?

04:16So say we all kind of opt into this and we want to make our businesses better,

04:19and we want to make our businesses safer.

04:21You have to build a model with some logic that makes

04:23sense as to what solving for. So who owns that?

04:26It depends on the situation, of course.

04:29But, you know, the way that we address it is there's a suite of privacy technologies,

04:33there's the analytics, but then there's the technically enforced governance.

04:36And so in any given collaboration,

04:38the different parties come together and they say, this is the outcome we want.

04:40This is how we want it, and then technically enforcing it is really the rub here, right?

04:45Because otherwise, how are you really protecting yourself?

04:48Right. Right. Well, I mean, that really is the question then when you think about

04:51data privacy and who's regulating this AI environment, Who is that?

04:55Is it, you know, we know that U has some framework.

04:58We're still working on that here in the US,

04:59and, of course, there's this proliferation because,

05:01you know, each individual state has its own set of standards.

05:05So is it almost like an SRO,

05:06like a self regulatory organization thing and just doing the right thing as an industry?

05:10Or where is this framework? Where does it exist?

05:13Well, for us, I mean, being in cross border payments all over the world,

05:16we basically have to go into every jurisdiction that applies to a payment, right?

05:19So, being able to open a corridor takes a long time,

05:21we have to investigate, how does it work?

05:23What other things we need to do?

05:24If you wanted me to do, you know,

05:26a payment for you in let's just say Australia,

05:28I would have to understand what's happened in Australia,

05:30what happens in the US? What happens on the way?

05:32But then once you have that,

05:34then you would basically have to follow what

05:35happens in jurisdictions as things go along, right?

05:38So it's a lot of work that goes before so that when you need the payment,

05:41which hopefully is near instant or instant in the best case.

05:44Then you're already ready, right?

05:46And that's the work where Karrie and I would spend a lot of time.

05:49But To say it's easier.

05:52This is a very difficult,

05:53complex problem we're all trying to work through.

05:56But in the EU where there is a regulatory framework where

05:58everybody abides by one thing in the US, like,

06:02we're managed at the state level,

06:04and different types of data,

06:05whether it's biometric data or whether it's PII data are managed differently.

06:09And so in those types of scenarios,

06:11even with some of the regulation on the hill,

06:13It is like this one, two, three layers of defense.

06:15Like, it's what am I doing to protect

06:17my own customers and the businesses that I'm working with?

06:20What are the people who are working with me doing?

06:23And do I trust them to your point of how do you do a KIB on them?

06:26And then do I trust the people I'm actually maybe

06:28giving information to in the process of payments.

06:31It's card processors or banks, et cetera.

06:34And you have to self govern at some degree until we do have a framework,

06:38I think, a global framework that makes more sense of this.

06:40And that's part of the challenge, too.

06:42When you self govern or self censor,

06:44you're actually reducing the value that you're getting out of the AI,

06:47you're censoring what inputs you're putting in,

06:50which censors what you're getting out and the value that you're getting from it.

06:53So really, you know,

06:55we see organizations doing this all the time.

06:57And so, obviously, there's regulation that you need to comply with.

06:59But everyone's kind of scared because you don't really understand

07:02where AI's going and where the data's coming from and how the data is collected are used.

07:06And so that's really where you need to protect yourself if you were.

07:10And, I mean, I think for a versa pay customer,

07:11to actually be able to trust you,

07:13you have to trust me if I was the one,

07:15you know, doing the for you.

07:16So we would have to spend a lot of time as

07:18an industry together to actually validate each other.

07:21And then when we're ready, then we build the environment in which we can

07:24then go ahead and that takes a lot of work. It does.

07:27I is in there. Then, of course,

07:29when the transaction happens there's millions of transactions all the time, right?

07:32Right. They are also screened,

07:34but that's a totally different So, Ron, let me ask you this.

07:36I mean, companies they're just rushing to spend billions of dollars on AI,

07:39and it's almost like, is there even a problem that needs to be solved?

07:42But the question is, you want to be first to go to the market?

07:44You want to establish revenue lines with this new technology,

07:47but then you also have to balance out the risk management side of it, too,

07:50when it comes to open access to your data or you know,

07:54never mind, you know, the cybersecurity implications around it, too.

07:57So how do you measure that potential return relative to the risk that exists?

08:01Well, to the point of what we were just discussing,

08:04I think the return is going to be limited if you're limiting yourself. Mm hm, right?

08:07And so I think that you'll see that the way that AI application vendors are going to

08:11differentiate themselves is how they can actually access and use sensitive data,

08:16proprietary data, and collaborate with their ecosystem with their partners.

08:19That's really the way to unlock data and unlock value.

08:22And so I think what you'll see is just like, you know,

08:25RSA unlocked Ecommerce, the world of

08:27Ecommerce by protecting credit card transactions online.

08:30You'll see privacy enhancing technologies unlock

08:33the world of AI and unlock the ROI that comes with it.

08:36Ronan, I do want to touch on something Jill asked about,

08:38are we solving a problem?

08:39Because I do think that all of us who have been in business,

08:42when J AI came to the mainstream,

08:45and it was part of everyday conversation.

08:47There was this rush to talk about it or

08:49put it on the end of what your value proposition is.

08:52And I think we're in a world now where you

08:55have to be using data to create value for yourself,

08:58for your customers, for your shareholders.

09:00At the same time, it's not going to solve every problem.

09:03Like, I look at the world that we're in right now where

09:05business business transactions, particularly in the field.

09:08It's an underserved market.

09:10They're still using legacy systems.

09:1133% of it is still managed by check, you know?

09:14And even though we're doing billions of transactions, you know, at the end of the day,

09:18you know, I look at is AI really going to solve

09:20the problem that this office of the CFO needs?

09:22Not necessarily, maybe over time, more sophisticatedly,

09:26but I do think that's an important point is,

09:28is it solving the problem or is it in search of a problem?

09:30And maybe we're just not ready in some areas.

09:33But I think that again,

09:34back to your point about the U and what happened there already,

09:37and with the whole data owner and the data data?

09:40That's the question. Exactly. But the data processor and the data owner.

09:42I mean, that is exactly what's defined in that law, right?

09:44Right. The point being that once you go digital,

09:46it's a lot easier to handle who actually can keep what at what point in time, right?

09:50Whereas the check, I think there's very few areas where

09:53the US technology wise is not ahead of Europe, but exactly in that area,

09:57actually the European economy is because a lot of it is actually digitized already,

10:01and that makes sure that we can take the data,

10:02put it into your model, use it the way we should,

10:05but protect it and actually remove it again.

10:06I think that's a very important Well,

10:09the EU looks at it as the consumer the data.

10:12I mean, it's a consumer driven act.

10:14That's that's sort of the That's the paradigm that the U laws. Yeah, I don't.

10:19Because when we look at President Biden's executive order, you had noted that,

10:22it seems as if it's more about the government and what it is allowed to do as it

10:27relates to privacy protection and combating other let's call it national security risks,

10:33but it doesn't appear that it's necessarily geared towards companies or towards people.

10:38It's more of a government action. Almost like the First Amendment.

10:41Yeah, you know, the executive order that you're

10:43referencing really is designed to protect the First Amendment,

10:45protect data privacy, and that starts with government.

10:48But, you know, there are federal AI acts, federal privacy act,

10:51and even state level laws coming through that will govern this,

10:56you know, on a consumer basis or even an industry basis,

10:58like in financial services.

11:00So I think the government in the US is kind of leading the way,

11:03actually and thinking about how you can leverage

11:05AI while still protecting our rights as citizens.

11:08And that's a really important benchmark for everyone I think so,

11:11too, if we work from a position of the citizen owns their data.

11:16And then when you look at a supply chain of how data moves

11:19and how it's created at what points who owns that.

11:23But I do think is a democracy in which we're in Data

11:27and AI is becoming part of the fabric of what our future holds.

11:32And we're talking a lot about financial crimes

11:34because we happen to be in the payment space.

11:36The reality is it's beyond financial crimes,

11:38like misappropriation of imagery and content and misuse and that's

11:43where it starts to get even more concerning for me of who has

11:45the right to create and then who has the right to use.

11:48And let's remember, we're all governed by licenses and all kinds

11:52of regulatory environments and

11:53authorities that actually wants to check what we've done is correct.

11:56And if something happens later,

11:57anti money laundering laws and so forth,

12:00we have to be able to prove who actually did transact,

12:03who was the receiver, who was the sender, where did it go?

12:05So there is something there that the whole law the sort

12:09of the way that the laws are built has to take that into account.

12:12I totally agree, it has to go into every part of the fabric. Otherwise, it won't work.

12:16Which is so interesting because when they included

12:18First Amendment rights within this act.

12:21My first thought was almost similar to yours is that this is kind of

12:23advancing the First Amendment into the 21st century,

12:27if you will, only because, you know,

12:28who does control our data.

12:29And then perhaps it's not us even sync.

12:32And the way we were talking about before,

12:33deep fakes, where what was Katy Perry was like,

12:35in multiple places that Scala wasn't the Mcal thought so.

12:40So it almost feels like, you know, there's that argument that some

12:42of the amendments in the Constitution could perhaps be outdated,

12:45but it sounds as if acts like this are incorporating some of that.

12:49Yeah, I think it's starting to,

12:51but I think it's just the beginning.

12:52Right. I think we would all agree to that.

12:54I mean, I was looking at the news a couple of months ago,

12:56and I think there was an incident in Hong Kong, if I recall,

12:59where this is where AI makes me nervous,

13:01where a financial controller was sitting in a Zoom call with his staff.

13:05Do you guys hear about this? And effectively,

13:08they were all deep fakes.

13:09And one was acting as the CFO.

13:11And the CFO instructed them to move $25 million, and he did.

13:16And so you start to look at digital imagery.

13:19The Katie Perry example is kind of fun.

13:22It creates a lot of social buzz.

13:24But something when you're moving money that is

13:26at the detriment of a business, that's pretty problematic.

13:28And so how do you govern that? Right?

13:30Because it's not the banks fault that the money's been moved. I did its job.

13:33You know, it's Second I got issued

13:35from the right person who was approved to actually move the money.

13:37Amy foreign minister was in a meeting with Defate. What's actually.

13:41So how can you be proactive about data security measures?

13:44Like, what are the best practices adhering to the framework that's in place?

13:47I mean, you're in financial services,

13:48so, of course, that's heavily regulated.

13:50You know, you do the best to adhere to that?

13:52What are some best practices?

13:53Well, I think, for us, I think we are we complete

13:57more than 99% of the payments we do that we asked to do,

14:00and that's quite unique,

14:01even if we didn't do complex, but we do that.

14:03I think the first thing you do is you build an environment that's actually safe to be in.

14:07And again, Kar and I would spend a lot of time before we actually start transacting,

14:12checking each other, making sure that networks that we put together.

14:14It could take months before you partner.

14:15It could take months. It could take months.

14:17But once we're there, we know that we actually have the right,

14:19you know, guard rails around what we're doing.

14:21We know that we're in the right environments.

14:22And then we start looking at transacting.

14:24Then what we do is we look at every transaction, we screen them.

14:27There's all kinds of regulation and stuff that we can do,

14:30we partner with companies like

14:31yourselves that can actually help us do that screening on the go.

14:35And then, of course, afterwards,

14:37we also then look at monitoring what actually happened,

14:39who did this, over 24 hours?

14:41Who did what kind of thing.

14:42So we always track and try to find out what's going on.

14:45But it's always about tracking the data and trying to break in.

14:48Well, I mean, it feels like Ronan is always going to be a game of playing chess.

14:51You're just always going to have to be thinking,

14:52step ahead and trying to stay ahead of the technology.

14:55Yeah, you always have to do that.

14:57And that's why you know,

14:58having a posture where you're protecting your data and still

15:02able to use it and starting with that as the starting point, right?

15:05How do I get to this data?

15:06But how do I use it in a way that keeps it secure,

15:09keeps it private, and obviously keeps it compliant?

15:12That's really where you need to start, right?

15:14Especially when the data is sensitive,

15:15like in the world of fraud or

15:17healthcare and life sciences or national security, like you mentioned.

15:20Care, you pointed out something interesting in your notes here that if

15:23we move forward with trust inclusivity,

15:26secure identity, ecosystems, that

15:28can really protect and promote economic growth. What do you mean by that?

15:32I think data is the core at how we can make

15:36different decisions in order to help people get the goods,

15:40the services, the assets they need.

15:41Like, you were talking about moving money into Miramar.

15:43You know, look at some of the tornadoes that are happening here in the US,

15:46that if we actually understand who somebody is, how they were impacted,

15:50and what they were entitled to, and we can do that

15:52through data and data that's harmonized,

15:54Jill, not data that is effectively sitting in all these different repositories,

15:58We can help create a better society.

16:00I think we can enable businesses to be more successful.

16:02We can enable citizens in order to live the lives that they want to live.

16:06Take, like, unemployment insurance, for example.

16:09Like, when COVID happened,

16:11there was hundreds of billions of dollars that were probably

16:14misappropriated and misused because businesses and individuals,

16:18pretended they were somebody they were not or they were impacted somebody or not?

16:21Think about that money went into

16:23other aspects of our society and the problems that we have to solve.

16:26And so that's where I think, like, trust, transparency,

16:29using information in the right ways,

16:31protecting it against the bad actors can actually spur economic growth.

16:35That's a great example that you, Thomas,

16:36you moved money to MinMar via the Red Cross in a day.

16:40How long would that have taken if these tools were not in place?

16:44Weeks. Yeah.

16:45Weeks have that kind of time when you're dealing with the humanitarian.

16:48That's what you have to deal with until these things are in place, right?

16:51And if you think about the world, if you think about anything outside

16:54the societies that we're used to being part of where financial infrastructure works,

16:58and so on, these possibilities assignments, right?

17:02I mean, you can include parts of the world that

17:03before did not have financial infrastructure,

17:06you can take people who actually have the right to thrive to get on with it,

17:09to do their business and get along with their lives. That's what we have.

17:12Does it ensure it gets into the right hands?

17:15I was just going to used to take three weeks?

17:18Moving it could take second.

17:19Does it get into the right people that take.

17:22That is where I think, again, coming back to what a society and what sort of

17:25an ecosystem is because what we do is we partner up with people that we can screen.

17:29We know how to get them there.

17:31We can get the money into Man Mar.

17:32We cannot control what happened inside Man Mar course.

17:35Right. But by screening the partners that we choose,

17:37by screening the corridors we use, that takes months.

17:40It takes actually months, but six,

17:42seven, eight months, if it's difficult.

17:44But once we have them, we know exactly where we put them.

17:47How they then disperse from there,

17:48of course, is out of our hands.

17:50But that's how the ecosystem has to work for us.

17:53But we can put the money into a country,

17:55you know, in near real time.

17:56Yeah. I mean, really is quite fascinating.

17:58I mean, if you think about it on a timeline scale,

18:00just how quickly this has evolved.

18:01Right. Absolutely. And you know,

18:04I want to go back to a point that are made

18:06in terms of the use of data and helping people, you know,

18:08whether they're in Myanmar or in Nebraska with the tornadoes,

18:11the data that we all need is actually very siloed,

18:14not just in our own organizations,

18:15but across different companies across different countries as well.

18:19And so in thinking about how we could do this,

18:21collaboration has to be a part of it.

18:24There's no way that any one organization can go it alone.

18:27The AI application vendors need the data.

18:30You guys need the data, right?

18:32And so the way that we can continue to innovate and continue this pace of

18:36growth that you just referenced is through this collaborative posture,

18:41but it also has to be secured the entire way through.

18:43And the disruption that's going to happen in this industry

18:45is exactly driven by that possibility.

18:47It's not just the need. It's the possibility that we can now do it better.

18:50We don't have as many areas of,

18:52you know, where faults can happen.

18:53We don't have so many places that things can go wrong,

18:56and at the same time, it's better service, right? So we get on with it.

18:58But I think it's incumbent upon us as leaders,

19:02whether you're in enterprises or companies or NGOs or in, like,

19:06large markets and diplomacy that we

19:09effectively have to educate people on how their information is being used.

19:13At the end of the day, like, 'cause I'm willing to give up

19:16information if I know my life is going to be better,

19:18or I can make different choices.

19:19And that is something that's incumbent upon all of us.

19:22That's also where the trade offs. Sorry. The trade offs

19:24can kind of go away with new technologies as well,

19:27where I know that my information is going to be used for some purpose,

19:30but I know that it's going to be done securely.

19:32That's exactly.

19:34Consent isn't enough anymore, right?

19:36You need the security because there's bad guys everywhere trying to get at this data.

19:40And we need to work together to make that much

19:42more safe because we get fast, safe and simple.

19:45If we can make it like that, it's going to work for the good guys,

19:47and hopeful can keep that guy.

19:48Alright. We'll leave it on a positive note. Appreciate everyone's insight.

19:51Thanks for joining us on T. Thanks.

19:52Thanks for joining me from Market site.

19:53I'm Joe Melenro Global Market reporter at NASDAC.

20:01I.