Subscribe
Share
Share
Embed
Welcome to The Payment Expert weekly podcast, brought to you by SBC Media. Each week we analyse the news driving the global payments industry forward; the innovation, the infrastructure, and everything that has to happen to make it all possible.
Louis (00:00.376)
Hello and welcome back to the Payment Expert podcast, your source for the latest news, insights and analysis on the payments industry. I'm Lewis Thompson, news editor here at Payment Expert and with me today, I'm delighted of course to be joined by Erica Stanford-Fizer at CMS UK. Erica, thank you very much for joining me today. No worries at all. You cover a...
Erica (00:23.549)
Cool, thank you for having me.
Louis (00:28.366)
of areas really when you think about how complex each are from guess crypto AI and how they sit across the ecosystem. Let's start off with a bit of a big talking point I should say sorry, which is stable coins at the minute. They process sort of more on chain value last year than visas network handled in total. So the use case is rapidly growing from where you sit in the advisory.
position advisory and I believe it's somewhere around 300 crypto firms. Forgive me if I'm wrong in suggesting that. How much of that is you seeing as sort of real payments activity versus trading and treasury movement?
Erica (01:09.765)
Yeah, the stablecoin volume is interesting. There's been a lot of hype with some of the stablecoin companies saying they do more payments and more transactions and then Visa and so forth. There's a couple stats in it. The last I heard was about, I think it's 13 point something trillion dollars of stablecoins in terms of total payments, but that includes a lot of automated and bought payments. So when you look at the adjusted volumes to take out the
automated bot payments that brings it down to just over three trillion dollars of payments a year and the 13 trillion in terms of transactions and volume is you know a big number compared to Visa and MasterCard but then when you take the volume that's just used for payments it's the last step I heard was 390 billion dollars so still a lot
still a lot of transactions, still a lot of volume, but not compared to the total transactions. there's a couple differences there. Stablecoins are used a lot in crypto trading. So that's a lot of the use cases, but the 390 is still a decent amount, bearing in mind it's still relatively new to be using stablecoins in payments and remittances in actual real world trades.
Louis (02:37.185)
Yeah, it's probably a growing use case too and different, I guess, markets have different rules. If you look at the US, the Genius Act bans, I guess, issuers from paying any yield, fixes it more as a payment instrument rather than a savings product. From what you're seeing, does stripping out yield in that market and perhaps in others to remove the main...
I guess reason why people would hold them as maybe the crypto industry would suggest.
Erica (03:08.249)
No, I think that there's two main reasons people hold and use stablecoins. The first is in crypto. It's the main settlement between different cryptocurrencies or used sort of to get out into fiat without actually having the costs and the complexity of getting out in fiat. So the main use at the moment is in crypto transactions where people don't want the yield, they just want the ease of getting in and out of assets. But the sort of the most important
real-world use is stable coins as payments and really they're intended to be digital money so you see a lot of these cases where remittances are paid in stable coins for example I've spent a lot of time in Latin America and there's a lot of use of stable coins where people rely on stable coins because some of the local currencies aren't stable and don't maintain their value and earlier before stable coins were a big thing
would go into Bitcoin and despite Bitcoin's volatility and Bitcoin is extremely volatile, it's less volatile than some national currencies and it was seen as a safe store value but now stable coins are the easiest way for a lot of people to hold cash in something like the US dollar and then people aren't worried about getting yield so much as they're worried more about is my local currency depreciating, is this going to be a stable store value, is my money more or less
in this asset or can I pay staff or can I pay contractors or can I pay gig workers or can I receive remittances in a stable form of value.
Louis (04:49.035)
Yeah, absolutely. I suppose from when we look at crypto in general, it maybe has a bit of a reputation still from banks and regulators or a perceived reputation that it's kind of crypto payments have a higher money laundering risk. There's associations. I think even last week we saw some crypto mining farm in Mexico, BCs, but it is more traceable than
Erica (05:05.403)
You
Louis (05:15.575)
cash in that sense, if we look at the money laundering side of things. Where do you think that perception comes from? Is it just the use cases? is it maybe a reeducation around actually if it's on an immutable ledger, it's a safer means of payment?
Erica (05:33.512)
So there's a few things. So in the early days of Bitcoin, in the early days of crypto, higher percentage of crypto transactions were used for crime. So sort of go back a while. The first couple years of Bitcoin, about 30 % of all Bitcoin transactions were related to crime. And mostly that was people buying drugs on the dark web because they thought Bitcoin was anonymous. Not true, not true at all. So going to now, less than one,
1 % of crypto transactions are related to crime in any way. So that's a lower percentage of crypto transactions related to crime than for fiat and also because crypto is so much smaller, it's also much less in volume related to crime. So now where it comes from that there's a few things you do now have incredibly powerful crypto analytics software and crypto analytics tools that allow you to trace crypto very well and even if very sophisticated criminals can
do lot of things to try and make crypto harder to trace these tools can still trace crypto so that the software is now really good so crypto is more traceable and every time you speak to someone in law enforcement they're always happier to trace crypto than they are to trace fiat or other assets it's easier to trace but there are a few risks so crypto you can move it faster it's seen as easier to launder for you know for good reasons because you can get hold of the assets you can move it
quickly you can cash it out in into cash you can use it for your casinos so it is easier to launder. The other factor with crypto is even though you can trace it that doesn't mean you can necessarily do anything when you found out where the money has gone to so if crypto has for example been stolen or it's proceeds of a scam or a crime or
has been involved in money laundering, you might be able to trace it to a crypto wallet and see, you know, know exactly which wallet that crypto is being held in. sometimes using other tools like IP addresses and other information available, they can find out who is holding that crypto. But that doesn't mean you can necessarily do something. So you can trace crypto quite often to somewhere like North Korea. And then there's not anything you can do about it unless they decide to move that crypto. Whereas, for example, if
Erica (07:56.336)
it's money that's stolen and you can trace it to a UK bank there's a better likelihood of being able to get that money back whereas you you don't always have that with crypto depending on the jurisdiction it ends up in.
Louis (08:08.013)
Yeah, I suppose it's the way it moves across jurisdictions so easily makes it harder in certain cases to obviously get that money back. Different rules and regulations apply in different markets, different types of states. Very interesting stuff, Erica. Thank you for your insights on that. Let's move into the AI discussion around, I suppose, linked to money laundering as well and scamming as we've been talking about. AI has made scams.
Erica (08:13.661)
Yeah.
Louis (08:35.775)
I guess cheaper to run fraud as a service has become almost quite a big industry in and of itself. You think of clone voices, deep fake video calls, fake platforms. When you look at payments fraud, what has sort of changed most do you think because of AI? know APP fraud is a big one in general, but is there anything that AI has particularly done for you in sort of perpetuating fraud?
Erica (09:03.965)
Yeah, well AI has made fraud a lot easier and it's really lowered the bar to entry. the thing with AI now, there's a few things. AI can be used in deep fakes and impersonations. So you can have videos or audios that look and sound exactly like the person they're imitating. And you can also use AI to create a whole evidence system for scams. So what the best criminals now do, they'll create a scam or
a form of crime.
they'll use because you can use AI to create content to create websites so easily and so quickly and so cheaply you can now create whole websites and evidence systems and reviews and testimonials and and everything that support the scams so that's another thing that AI has done we talk a lot you know in in the real sort of non-criminal world there's a lot of AI noise and AI creates a lot of just content and slop everywhere it's the same in in crime so you can create so much back
grounds of noise for for crime and on the dark web you've got the whole crime as a service ecosystem where you've got scams as a service tools being sold and money laundering as a service tools being sold and and that whole ecosystem combined with the tools available mean that now nobody really needs to have any technical skill or even any Linguistical skill or any social engineering skill to be able to scam and to create very very convincing ecosystems and it's just
removed a lot of the ways that you might have traditionally known that something is a scam. It's removed all of those because you can now use AI to make anything sound like you or sound like me or to message in any chosen way in any chosen language or dialect. So it makes it a lot harder now to know what is real and what isn't real.
Louis (10:57.197)
Yeah, it's very tricky time as well, I guess, and also the origins of those scammers too can be across jurisdictions and these kind of, guess, international cooperation in certain aspects to catch them. organisations across various industries are adopting AI as well for their protection and...
Some suggest there's a bit of an arms race in terms of who's got the better use of AI, is it the fraudsters, it them? But then I suppose in protection terms, a lot of firms have got it right. And then in those cases, you see fraudsters maybe move to attack the consumer rather than direct attack on any organization or company.
Erica (11:24.317)
Yeah.
Erica (11:38.406)
Yeah, there's a few things around that. So criminals tend to win in terms of technology because they're not limited by things like law or by data protection. they also share information a lot more quickly and a lot more freely with each other, which is really useful for them. So some companies are doing really well in terms of preventing against cyber crime, in the terms of a crime. And it's not so much that the ones that are doing really well are doing anything sort of better or more sophisticated than
criminals it's more that they're raising the bar to attack them and so from the criminal point of view if you've got some companies that it's harder to access or harder to get into then they'll just go and find one that's a lot easier to attack so one incentive for companies to sort of use cyber security properly and also to use AI defensively is you want to be the one that's harder to get into
so that the criminals don't aim to attack you, they'll find an easier prey.
Louis (12:42.925)
Yeah, totally. And I suppose in banks, they're leaning on AI to block certain suspicious payments and legitimate transactions get frozen in the process sometimes, perhaps. And I suppose there's a new line in banking at the minute, isn't there, where it's protecting customers versus, if you overdo it, you might be locking them out of their own money in certain cases.
Erica (12:53.937)
Yeah.
Erica (13:07.045)
Yeah, it's a really hard balance because in some cases, and we see it as well, we have lots of banks and companies as clients, in some cases a bank knows 100 % with absolute certainty that the address or wallet or account that one of their customers wants to send money to, they know 100 % certainty that that is a scam. And you hear of cases where the customer then sues the bank or whatever because they wouldn't let the bank send
money to them even though the bank knows 100 % that it's scam some people just don't want to know or don't believe it or so. So you've got the full spectrum of it where the bank sometimes really does know best and really is acting in the best interests of its people and know banks also have to reimburse for it so they're not incentivized in any way to let customers send money to scams but then sometimes it does go too far and block transactions. There is a balance where often scams
and frauds, they just rely on pressing, putting pressure or putting urgency onto people. So sometimes all that is needed is just adding a delay and adding some friction and asking more questions. Like, this is a scam.
you're trying to send money to. Do you know this person? And sometimes just asking questions and adding delay is all it takes. And in some cases I think there is an argument for freezing transactions when they know that it's best. But there is also a problem that anything, not just banks but platforms, the more security you offer to your users also means the more surveillance on your users and the less privacy and freedom you give them. And there's no clear
answer as to what's best on that so that's a balance that every company has to look into.
Louis (14:58.699)
Yeah, it is a fine line with that. also the banks themselves get, I mean, they're hyper-regulated more so than maybe, you know, a platform where a deep fake attack could come from and they have to protect obviously their, know, their customer just as, you more so than probably any other industry would have to protect if that makes sense. So if someone were to...
you know, get their accounts drained by an AI driven scam or, you know, a deep fake of their bank, a cloned voice. Where do you think that loss should really, really sit? Obviously, you know, is it with the customer for being duped? it the bank as it kind of is or the platform that it's scammed on? A lot of it coming from social media, I imagine.
Erica (15:42.237)
you
Erica (15:46.782)
a lot is from social media. So I mean, in some cases, and we definitely see cases where it's the person, the individual should be the one to pay. There's definitely cases where people get greedy and send money to scams, knowing that they will get reimbursed if they don't get the money. And, you know, they see it as gamble, know, either we'll get this, you know, guaranteed doubles return or whatever the promise is offering, or we'll get our money back. And you see that a lot and also in the UK.
the reimbursement rules came in, fraud in the UK has increased, not in every country but in the UK and there's a direct correlation between the two. So I definitely don't think it should be the bank in every circumstance, that wouldn't be fair at all. And there are social media platforms that profit from scams and that profit from fraudulent ads. We don't need to go into naming anywhere any here, it's very well
documented and that's a real problem and there are social media platforms that have adverts that are fraudulent or that have accounts that are known to be fraudulent and even if they're reported they don't do anything about those and the people controlling those accounts and even sometimes the same adverts stay on and keep scamming and keep scamming and that's a massive problem so I definitely think that there should be an awful lot more of owners on the social media platforms and not just social media platforms
advertising platforms and that might be messaging platforms where these scams are able to originate and at the moment there just isn't enough done when a scam is found and reported between reporting that and blocking that. I think that's a massive, massive barrier and a massive problem.
Louis (17:36.95)
Yeah, and you're not the only one who's told me that, It seems to be quite a running theme that more needs to be done on that front to block those kind of malicious ads and things. But I know at CMS, you've been sort of working or you're involved in ethical AI frameworks that you're helping kind of develop. Tell us a bit about those, the kind of use cases where AI is being deployed, where maybe an ethical framework could come in and sort of help.
Erica (17:40.002)
Erica (17:55.065)
Yes.
Louis (18:06.721)
have it be established in the right way.
Erica (18:09.467)
Yeah, it's a really interesting field and I actually just finished doing a masters in that in data and AI ethics, which is all about all of the risks in AI, everything that could possibly go wrong about AI and data and what you can do about it. And on the ethical side, there's so many different angles and some companies have different priorities. In some cases, AI uses a lot of water. It uses a lot of energy. has massive environmental impacts. The materials
that go into creating the of the physical devices used in AI use a lot of rare earth minerals and rare earth metals. They tend to come from areas where there are high levels of exploitation. There might be unemployed labor, there might be child labor, there's also massive pollution and environmental damage in those areas where those materials tend to come from. So that's one problem. There's all the different angles around AI. In some cases, it's how AI is
trained. There's a lot of people whose job it is to look at the content that either AI creates or the content that AI is trained on and to do things like mark what's wrong or mark what's distressing. And in some cases you hear awful reports about the content that they have to look at day in, day out, day in, day out. And these are people who often paid something like a dollar a day and whose jobs are just ended just at the whim of companies because they want to move their labor to a cheap place or so forth. So there's
There's that issue. And then there's other issues in terms of AI might create mistakes, it might create bias, it might have vulnerabilities in it that lead it to be able to be manipulated or it might lead the data in it to be poisoned so that even the people creating it or the users, it might happen so slowly that you don't really see it immediately, but then that might end up causing problems. There's also issues when AI is used to make decisions without
human review and sometimes those decisions aren't fair or might target a certain group and there might be certain biases brought into that or certain forms of discrimination that it might be that the AI has been trained on that knowingly but often not and it might be that it's just picked up its own biases or it might be that the people training had their own biases that they weren't necessarily aware of or it's just the AI has been exposed to the internet
Erica (20:37.591)
and you know there's problems on the internet and not everything in the world is fair so it's learnt its own ways to discriminate so there's a lot of potential risks and a lot of potential ethical risks and also legal risk all of which have ways to mitigate them it really just depends on what specific risk you're looking at then there's different ways of doing that it's a total minefield
Louis (21:03.263)
Yeah, I suppose it's diagnosed the problem and then there's the fix for a certain issue in certain ways. Really fascinating stuff, Eric, obviously, the way AI is being discussed today. It's obviously a very important issue for all being one of them, as we've discussed today. Just lastly then, from me, if there's one thing you would recommend perhaps a payment firm to change in the next...
year when it comes to AI-enabled fraud to get ahead of that fraud, what would it be?
Erica (21:34.669)
one one thing I think the main thing I would say would be education and I know that's totally cliche and I know it's been said to death but the one thing is AI is changing so quickly the type of frauds are changing so quickly the criminal networks running them are learning and developing them way
quickly than legitimate companies can keep up. yes, there's a lot that companies can do to try and keep up and try and employ cybersecurity and so forth. But in terms of preventing fraud, a lot of people just don't know the risks of, for example, clicking on a link where you don't necessarily know where it's come from or things like checking who actually sent an email because the address might be spoofed or if somebody, a recruiter sends you a PDF on LinkedIn about a new job offer,
it's a recruiter sending you a job offer or maybe it's a North Korean hacker trying to get you to download malware to get into your company device or to try and get into your crypto wallet or something like that. I think the main thing that if it's just one thing educate your users into all of the potential risks but even you can't educate on all of the potential risks I think just educate people to be more cautious and to just not trust anything inherently.
you
Louis (22:55.917)
Yeah, sound advice, think Erica and hopefully our viewers and listeners can can take some positive steps from that. Thank you very much for joining me today. That is unfortunately all we've got time for. If you've been watching and listening, as I've just said, and you're not already subscribed to the payment expert podcast, make sure to subscribe wherever you do get your podcast as we've got plenty more insights coming up over the next few weeks and months.
And for the latest news as it happens, can head over to paymentexpert.com. We'll see you all next time.