Fraud Prevention in Crypto Payments

Hi everyone. Uh hope you're all well. It's quite hot here today. Just landed this um this morning. Um I'm Ulissiz.

I run global revenue at Hyper Native. And today I'm going to talk about uh fraud prevention crypto payments. Now to give you an understanding of maybe the magnitude of the the problem, uh I'm going to share a bit of data on this. Uh in 2023, we have identified 1.8 8 billion being stolen through fishing, scamming and fraud um in web 3.

In 2024, this number went up to uh 2.2 billion and we expect this number to go up uh even further next year or this year unless uh you know major measures are taken. Now, as you can see, we also split this across uh a set of uh exchanges and the largest exchanges that you see are obviously mostly impacted by this. Now make no mistake about it. Uh these exchanges are spending millions of dollars in compliance and fraud prevention solutions.

Uh so they're definitely you know interested in in stopping this. But the truth is that the majority if not all the solutions in the space today are not providing uh the right measures to stop this and to prevent it altogether. Now this is mostly due to uh four uh reasons. um attackers uh criminals are becoming more and more sophisticated and they're using more techniques to essentially uh bypass um what uh traditional mail vendors uh are able to to cover today. So one example is the the usage of uh benign addresses.

So essentially you know if I am uh the user of an exchange and I would uh be you know fished uh or would be facing a scam and I would be asked to you know make a withdrawals to an address if this address doesn't have a direct exposure to listed activities in most cases um traditional ML vendors would not able to to see that this is connected to a much larger uh fraud and scamming ring. Another thing that we have identified is that uh there's been uh misuses or overusage of the word uh real time. Most companies uh that are providing AML uh services uh and I know a bit about this because I spent six years uh at CH analysis. So I know a bit about this topic uh do not have real time um capabilities. they have delays of minutes, hours, sometimes days, which is lifetimes in our industry.

Another issue that we have identified is that there isn't a proper connection between uh off-chain and off-chain uh detections and this causes further delays. Uh but it's also due to the fact that all of the detections often are coming from reactive approaches. I'm going to talk about this in a in a second. And then the last component which I think most companies are trying to to catch up on is the usage of essentially AI to accelerate but also make pattern recognition more advanced and this is one of the major gaps that we've seen uh as well. Now this is an example of uh you know a pig butchering scam which essentially happened on Coinbase.

And the problem that you see here, this is our platform is that essentially you know an an exchange might be able to identify reactively uh that one of their users is being scammed. But by the time that they essentially block this interaction, this user could have been scammed for hundreds of thousands of dollars over a course of one week, one one month. for example, and this is exactly what happened here with Coinbase. Now, another important thing is that now exchanges, VASPs uh are becoming liable for these inter uh for these for these um essentially scams. So, essentially the bank uh and the exchange needs to pay back uh the user.

So, this becomes much more relevant than for exchanges to have the right tools in place. Now what we have built um at Hypernative is essentially a solution that allows uh the fraud team, the compliance team to have uh at their disposal an investigation automated investigation tool. Essentially what you see here is a graph that was built in seconds that essentially built the entire fraud ring of the scam of this uh you know like uh fraud that happened on uh on on different exchanges. This is an example from Coinbase. And the beauty of it is that you're essentially able to uh create this graph automatically, but you're also able to uh essentially identify the key actors within this ring uh in a in a matter of seconds.

So this is a very unique innovation that uh I think that all companies interacting in payments should have. This goes back to what I was mentioning earlier. There are a lot of scammers who are very active and they're active until they're reported. And this process is very manual. So most exchanges are getting reports manually about uh scams and frauds.

And this makes it very difficult to act in real time which is what is needed. Another uh example is one of curve where obviously there was a front end attack uh that was pushing users uh to essentially exposing them to to fishing. Uh with our platform we have identified this in real time and we could have uh prevented this. And the idea is that um essentially you know what you have today are platforms that are using reactive approaches uh platforms that do not have interactions uh or good labeling of both offchain and on and onchain uh interactions. And what we do at Apprentive is essentially uh labeling all the smart contracts and transactions operating all the uh chains that we that we cover.

We identify risk insights. We share them with our customers and then we essentially enable them to take actions. So pause a protocol uh or unwind the position in real time and this has saved billions of dollars in the last uh 24 months. Now if we look at payments you know obviously this could reduce chargebacks uh this could reduce exposure to fraud, fishing and scamming and the the most important thing uh is probably that all this can be done in a holistic way. So this is not simply covering a front- end type of attack uh or a fishing attack or a fraud attack.

This is a platform that allows you to essentially cover all these threats at once. And this is what is missing today uh in the web free uh space. And that's me. That's it. Thank you.