Demystifying AI Regulation, Innovation & the Future of Financial Services with Colin Payne

Speaker A: Welcome to Dave and Dom Demystify, a fintech futures podcast helping make sense of the world of fintech and digital finance. Please sit back and listen as the two Ds take a subject and discuss it to make it clearer and easier to understand. Demystify.

Speaker B: Colin, welcome to the podcast. It's brilliant to have you join us from your man cave. I'm, um, sort of looking at Space Invaders behind you with lots of envy,

Speaker A: but, well, some of us are, ah, old school. What can I say?

Speaker B: Well, it sort of takes me all the way back to when I did swimming lessons in the central pool in Reading, and my prize for doing swimming was always a bag of Monster Munch and a game of Space Invaders.

Speaker A: Oh my God. Ah, Monster Munch. I'm missing that. I'm going to have to bring that back

Speaker B: anyway. So welcome to the podcast. For those that don't know you, could you do a brief introduction to yourself and the FCA as well?

Speaker A: Yeah, of course. It's great to see you guys and, um, thank you for inviting me. My name is Colin Payne. I'm the head of innovation at the Financial Conduct Authority. I also am the chair of gfin, which is the Global Financial Innovation Network, which is a, uh, another body. So most people know the fca, it's the conduct regulator for the uk and the chief inside of the work that I do is a group of 100 plus regulatory bodies and we try to convene new types of conversations around the international forums. But majority of that work is experimental and practical. So we often gather ourselves to explore new technologies, new subjects, new ideas, new new frameworks and try to experiment in a practical sense. So we use the platforms we have and we work together. So I guess my job is to make sure the regulator is really close to that emerging technology. We have to identify those technologies that are going to reshape financial services. So that's often a bit of a predictive piece of work. We're looking to work with technology providers and those who are on the cutting edge. So a lot of my work is that kind of activity, but of course the majority is helping firms as they face into that new technology. So we, uh, want to help them do that responsibly. So the way we do that is to, to use some of our tools that we have in the fca, things like sandboxes that are now quite well known, replicated around the world. But of course I bring a lot of experience from the delivery world, so delivering digital products and services. I've worked in a number of different sectors, but probably in the last 15, 20 years. When I first met Darm, I think was about 15 years or so ago, we were doing digital transformation back then. So I guess we're still in the digital transformation world. It's just become a little bit faster, a little bit more intense and maybe a little bit more complex. I think there's change all around us at the moment now. So I think back in the day we were pioneers. Now we're part of a gold rush if you like, where everyone is looking to exploit the technologies. But of course I love working for a regulator. Allows me to have a grandstand seat in what's going on. And really the reason I came into regulation was to bring that experience and deploy it in a way that could really benefit consumers, systems, markets and try to make a difference to the world. So a little bit purpose driven as well. That's where I am, that's what I'm doing.

Speaker C: Fantastic. It's funny going back to like the digital transformation stuff, Colin. The three of us have been through kind of all of those cycles, but I have this feeling and I just wondered whether you kind of feel the same thing. But we've seen like the revolution that the PC brought about, then the Internet and then mobile. Have we seen anything like the level of transformation that we're about to see? Do you think this is just as big as the Internet or maybe less impactful or the current kind of climate that has quantum and it has agentic AI and generative AI and tokenization that seems to be so much going on. Has there been a point in your career where there's been so much going on at the same time?

Speaker A: I don't think so. I look back in the 90s. I mean of course the Internet as it flourished in the 90s was an extraordinary time. But I don't think we understood the impact and the potential of what was going before us. I mean, I don't know about you, but certainly when I kind of was a passenger on that period, I remember, I mean, it's funny, we meet people along the road, don't we? I mean, and this is how we develop our thinking. I met Tim BERNERS LEE In 1991, I remember very well and I didn't really think much of it, to be honest. Now I look back as a slightly, uh, more wise individual perhaps, and I think, crikey, that was a real privilege to be as close as I was to some of those early developments. And I'm sure I took that on board. And I've always Loved attempting to keep myself on the cutting edge. And I'm incredibly lucky to have been at the right place at the right time, one or two areas of my career. And some, you know, people say you make your own luck. But I do think sometimes it's just that, uh, it's happening to be at the right place at the right time. And, um, with that meeting, I guess I looked at this and, um, like everyone, I was like, well, this is really interesting. Did I have the foresight? Did I really understand the impact? I think because I went through it and you might know, guys, I'm a massive, uh, musical fan. You can see musician stuff behind me. I remember watching Bowie talk about the Internet in around 1998, 99.

Speaker C: And digital money.

Speaker A: And digital money. Well, exactly. Or Bowie. Net. It was a decade ahead of his time again, so there's someone who really understood it. And I'm not going to be pretending that I'm anywhere near that clear, uh, having that clarity and understanding what's ahead. Some people are blessed with that as an absolute gift. But I, on the other hand, have to spend my 99% of sweat understanding it rather than have that 1% of inspiration, you know, dominate. But I do think now, as you say, D, we've got that convergence of an extraordinary number of factors that. Well, my prediction is they do interlace and they do have an aggregating effect. So when you consider tokenized assets and the move to properly oriented blockchain economics, you look at the possibility of decentralized finance, you look at these new business models, you look at the technology advances being made every three months. It's astounding. I mean, look back three years and what we see now is we have to realize this is like magic to us three years ago. It's just incredible. We sit on our phones and we can get answers to everything we need. And I know that was the same on the Internet with a bit of searching, but, you know, we know the frontier models are really quite extraordinary now at what they do. So I think this is different territory again. I guess the difference for me now is I'm older and wiser, so I'm appreciating it more. And I'm also really keen to be inside it, being present in it whilst it's happening, rather than observing and it happening to me. And I think that's the same for the uk. That's my job, is to be present for the uk, inside that change and position the UK financial services at the best possible moment in that transition.

Speaker B: It's really interesting, because I guess Darm and I have lived sort of almost equally long as you in terms of the change. And it's sort of talk about exponential curves in terms of the speed of change. Like, I guess before the Internet there was books, and books had been around for a long time. But now we're in a stage where, like, I can't keep up with what's going on. It's almost impossible. I have to ask the models themselves to kind of keep me abreast as to what's going on. But I think for a regulator, that must be a massive challenge. Just the speed of change, like, particularly, I'll probably say a few unfair things and please push back. But the regulator itself may be more used to things happening at a slower pace. How do you kind of respond to that acceleration as a body?

Speaker A: Well, I think you can choose as a regulator how you lean into that. I think I'm fortunate in that I joined the FCA a year and a half ago now, just a little bit over a year and a half ago, and, um, it so happened to coincide with my coming in that there was a more deliberate mandate to accept. This is a reality, that speed of change will challenge regulatory process. That was a reality. You can choose to ignore that reality and you can sit and regulate from ivory tower. That's one option. You can lean into it a little bit, or you can really think of your strategy as something of a operational advantage. If you get it right, it goes back to that point we mentioned earlier. It's being present within that technology change. If you are, and you're keen on, um, making sure that you can do your best to match the change that's going on around you, then you have to think differently as a regulator. It's not going to be the same process. You're not going to be able to create a rulebook for every new tool, for example. So we've been very clear with AI that we're not going to write sets of rules because AI is changing rapidly and it's going to be a race that you lose if you try to do that. If you try, your process will take longer, as you say, to write the rules, to, uh, consult to the legislative process just takes longer. So the FCA keeps pace in a slightly different way. We accept the technology moves fast, and we might talk a little bit later on about tech neutral into tech positive. We might cover that. That was an important strategic, I suppose, strategic view that we were taking. We were looking at it and thinking, okay, well, how do you view Technology as a core in everything we do. So there is a move in the strategy, but before that we're principles based anyway as a regulator, we're not a rules based regulator. So we're focused on the outcomes for consumers and for the system. So we tell firms what good looks like. We keep a very close eye on how consumers are being treated by firms, but we give them room to work with any technology to ensure that it fits that outcome. So that approach really ages well. Consumer duty, yes, it needs work, yes, it needs refreshing, but it covers a lot of this activity and allows you the flexibility to deploy technology in whichever way as a firm you think is going to meet that outcome and objective. It doesn't matter whether it's a spreadsheet or whether it's a large language model. Both are equally, uh, treated from our point of view because the outcome is what we're focused on. So I guess that's really important. We also get really close to the technology. I said that part of my job is trying to do what you're doing, Dave, which is to uh, meet the technology, to understand what's coming down the road. And I do have the privilege honestly of being in a position where the FCA is afforded great access to high performing technologies. So we have labs, we have relationships with providers of technology. I've lent into that quite heavily. I've partnered with Nvidia, with Anthropic and with others who allow us access to really understand where this is going. There's nothing better than having a seat next to those who are really delivering in order to understand what the impact can be. And that's a win win for the technology and for the financial service. Because of course we all ultimately want this to be a good, solid system. No one wants it to be a competitive system which then loses focus on the consumer. That's a losing situation. So I get a great benefit from being in a macro position. I can go and visit partners, supporters, we can convene almost unlike any regulator on the planet. We can bring people together. And of course my, I mentioned gfin, that gives me the global perspective so I can work with other regulators, whether that be in Singapore, Hong Kong, Dubai. It's something Dharm and I have often spoken about online. I get a really front seat view in what other regulators are doing as well. So I can be quite confident when I say actually the UK regulation position is strong, it's equivalent elsewhere. Yes, there are certainly there are races that are being had in certain technologies and certain products, but we take it very Seriously, to ensure that we deploy these things in a safe and responsible way. So that's kind of where regulation for me sits. Yes, it's never going to beat the technology that it's working with, but it can get a great perspective and it can make great decisions based on working with industry. And that's what we try to do.

Speaker C: I mean they've touched on a great point which is about the speed of evolution. I guess the other thing that's kind of complex for not only just the regulators to manage, but also your customers, the banks. Right. Is the complexity. Because it's okay when we can deal with agentic AI on its own. It's okay when we hear a separate team looking at Quantum, it's okay for another team to look at tokenization. But are uh, you as a regulator and can you say whether uh, your interactions with banks also starts to look at all of this stuff together? Is there an overarching thing that kind of says we need to look at the combinations of these things because Quantum could accelerate AI even further and stuff like that. What's your views on this convergence?

Speaker A: Yeah, it's a great convergence is something we've spoken about for a few years, isn't it? I mean we've always predicted the converging of technologies would lead to something quite extraordinary. And I guess we take it to the. We still operate in that principles base, so we're still concerned about the outcome. So if that's the umbrella that we talked about already, underneath that umbrella, yeah, we've got work streams that focus on areas. Now we're looking at data, for example, and the interaction with AI. It's pretty obvious to anyone who works with this technology that strong data, uh, foundations are required for outcomes of AI. You can't start talking about deploying an agentic AI with shoddy data. You're going to end up with, um, I won't use the phrase, but obviously you're going to have something poor quality coming out the other end. So we obviously have a very strong data analytics capability. Now we've defined that as being focused on smart data. So we start to think about data, uh, in the context of financial services of the future. So open banking was deployed. I think all three of us probably had some scars on our backs from open banking. It was a tough deployment. Right. I was working at the bank side at those times. So I bring some of that experience to the regulator, uh, view. So now when we start talking about open finance, of course I can start to deploy some thinking around that, which is based on clear infrastructure, looking at the available technologies with which to deploy open finance capability. And of course I come at it from a use case perspective, you know, because I'm a, I'm an old product guy. So of course I'm going to be asking the question what on earth is the benefit of open finance to the normal consumer? What difference is it going to make to Mrs. Smith up in Liverpool who really has a bank account with one bank? What does she care? How could it make a difference to her life or his life or whatever? So I'm always thinking that way. So it's not probably surprising to anyone that knows me. I was supportive of spinning up a smart data accelerator which could be use case focused. We would work with the industry. We've got an initiative called Prism which looks with the industry at the right use cases for open finance. And unlike open banking, what we're saying to industry is look before regulation comes down on you because it is going to come, it is coming. But how about we work on the benefits to the consumers and to your firms? So rather than it being a mandate which is just disruptive, could we make it a constructive business opportunity? Can we make it a constructive consumer outcome? And now that's a pretty different way of thinking about things for industry. But it talks about convergence and I pull that thinking into the AI lab, do the same thing there. How do we work across use cases which we're going to inform the future of Agent E Commerce? You mentioned agents Darman. Dave, we've got to think about how this works in the future. When we give the right of execution to machines. What does that mean for supervisors? What does that mean for the firms, among the agents? Where does the line get drawn between them? So we're thinking about all the time now. Quantum you mentioned is really interesting. I think the first thing probably is to switch around. I think you mentioned darn, um, that Quantum could accelerate AI, which is certainly a possibility. I'm probably not as with quantum as probably I would need to be to understand that option. But what I do know is AI is accelerating quantum. So the impact of AI modeling and the new algorithms that are being developed are uh, extraordinary. And I get a bit of a soft spot for mathematics and algorithms because I don't know whether you know this, but I started early in my career developing algorithms as a trader in the late 80s. Yeah, I was a trader. I used very basic compute. It literally was writing in basic and Fortran and then creating some very basic maths algorithms to try and give me a bit of an edge in what I was doing on the trading floor. And that actually developed into work I did with ubs, which was more high frequency work. So there has been decades of some of this work already where those who are at the frontier have explored the convergence of some of these ideas. And I think now with Quantum, I mean people say, well, it's always been 10 years away. Our view is now we are very focused on the potential impacts of Quantum, um, in the next five to 10 years. So we're very clear on the roadmap. But it is informed by other activity that we do. So when we see that rapid increase in deployment of AI, of course we map that to the potential for Quantum and we quite carefully what we say to industry. It's time to get ready for uh, this new wave of technology. Now readiness is very, very important for our industry now. I think both of you, I'm sure, have done as many transformation projects as I have in it. You know how long it takes and how complex it is and how difficult it is to get the budgets for some of these things unless the regulator says it needs to happen. So whilst we might have a bit of a joke about open banking, it was the mandate that pushed that and pushed us ahead of our, uh, European colleagues. Using their same framework. We did develop something on top which was a distinct differentiator. And I suspect that this is what counts in conversion technologies. It's being identifying how these things are going to work in the future and essentially speaking to the industry with a tone which gives them guidance, gives them that kind of opportunity to change, and then gives them a clear guidance as to what they need to do and when they need to do it. And I think us and the bank have done that quite clearly. Our current reflection of this is there could be some advancement of that as well. So we are looking closely at how those activities of the firms that are developing Quantum solutions. Are they actually speeding up now? Are they deploying AI In a way? Should I be looking at Quantum Lab for example? Because that could be a reality in the next couple of years. So I guess that's where I'm at.

Speaker B: Yeah, I mean it's a fascinating because, you know, Q Day has, there's the sort of existential risk of Q Day and I guess that's not just about the future risk, it's about the kind of data gathering risk as well, which sort of exists at the moment. I think if companies aren't looking at sort of uh, upscaling their cryptography and things in respect of what might be coming. They're putting themselves at risk now, even though Qday may be five to 10 years out. So it's interesting that. And actually, for me, it's brilliant to hear just how involved you are, because my background, like while Dharmesh and you've been in the technology, I've been doing the user experience and design for the last 30 years in finance. And it was always compliance which held us back. I felt that's probably deeply unfair. And, you know, it got to the point where I was like, God, we just need compliance in the ring from the beginning so they understand what's going on. And that became something that we kind of insisted on. And actually then compliance people put became real kind of partners to us in terms of the work we're doing. But now it's sort of compliance and regulation has to be slightly out in front of what's going on because I guess if you really think about the future and the risk to consumers, there's a huge number of unknown unknowns. But if we don't take some of those risks, then we're probably, as a kind of country, we're gonna fall behind. So I guess a question for you is, how do you sort of balance the risk of these unknown nerms with that kind of growth agenda? Because UK plc, there is no reason why, with our stature and finance, there's no reason why we can't leave the world in terms of use of AI and quantum and all these other things from a consumer point of view.

Speaker A: Yeah, I agree. I think it's fair to say that regulators, they have waves of risk aversion depending on geopolitical and other circumstances, economic circumstances, understandably, you will have periods of history where there's a great focus on a risk. Probably, maybe. Unfortunately for all of us, we went through the late 2000s and the financial crisis and it was pretty. You were right on the edge of a catastrophic failure. I think all of us that were there saw that and saw it playing out and we were almost powerless. I don't know about you guys, but I was in banking and it just seemed like the dominoes were falling. What was happening, this shouldn't have happened in our world, but out of that was born a great new opportunity. With fintech, with disruption, with challengers. The UK really did well. Cma, uh, uh, FCA being born out of that. Suddenly we had this amazing period, unprecedented period of activity and growth. And look at the names that have been born out of that time. The Revoluts, the Wise, the Starlings, the Monzos. We punched so heavily above our weight during that period. But, you know, you then had Brexit, you then had Covid. These aren't surprises that we then had a UK which had to reset itself politically, economically, from a social perspective. But I think what's happened in the last couple of years is there's been a changing of the vibe, the mojo. Okay, what do we need to do? As you say, Dave, we can be in a great position. We may not be the number one in AI in the world, but we certainly can be the number one value build using AI in fintech, for example. We are brilliant at those experiences that you really have been on the cutting edge of for 30 years. There is no question that the UK has brilliant design shops. Jony. I've was responsible for most of the brilliance of Apple in the time that I knew them and going through the major product surge. Obviously Jobs had the vision, but I've executed the experience. And it was his spiritual compadre through that whole period. And I feel like the UK can be that spiritual compadre, can be that creative, can be that trusted, brilliant experiential advisor. We are the rock and roll nation. We still do that stuff brilliantly. So I think probably I'm waxing a bit lyrical in terms of the creative, but we are really good at it. If you balance that with how brilliant we are at the trusted professional services and the depth and breadth of our financial services sector. Well, I still would probably maintain there's very few places in the world that touch us. You can look at all of these really proliferating areas still. There's something very special about the uk, I think London is definitely a centre, but also Manchester, Liverpool, Leeds, Edinburgh, we've got some amazing financial centers and some amazing technology capabilities. So when it goes back to risk, which was the point of your question before I got slightly sidetracked on my creative side. I guess that's why our strategic shift was last year. We talk publicly about rebalancing risk, we talk publicly about the fact we should be considering the risk of not doing something. So the true creative mindset is quite disciplined. It will consider the risk of doing something, but it will balance that with the risk of not doing something. And, um, that's the kind of interesting world I live in where I can start to look at growth opportunity and balance that with the risk of doing those things and moving into a market. And you guys know I've been quite close to crypto assets and, um, we've had one or two discussions around that in Terms of, you know, is the UK lagging behind? Are we going to miss out to another jurisdiction? Well, I won't name any names, but there are many we could choose from in that conversation. And look, here we are eight days before the US goes into its holiday recess prior to clarity and prior to the Genius act going through. So there is a, ah, certain extent where I can look back and say, well, Our regime's published 30th of June. Thank you very m. Much done and dusted, goes live very shortly, probably six months ahead of most others now, if not more. And I might say, and I'm sure my colleagues in policy wouldn't mind me saying, I think it's probably quite a lot better than many other frameworks because it really has taken its time to develop. But most importantly, spoken to industry. Now, I'm not saying we've done everything industry wants, we haven't, because we can't. I mean that's not necessarily what we're trying to do. But we have listened, we have engaged and, um, we've actually changed positioning on certain elements as the consultation has gone. It's been incredibly transparent. I think that's the other thing. This is not a behind closed doors thing, it's an engagement, it's a discussion challenge. Let's move forward. So that's where risk is really, for me, played out. Uh, in reality it's okay, let's talk about the risks. We think there's a risk here, industry. You prove to us there isn't. Prove it. I have to say prove it is probably the most powerful two words, uh, that I use currently to anyone that comes to me, because most people are extremely positive about their new technology or their new product or their new service or their new experience. Dave. It's like it's going to be the solution that the world has been looking for. Nothing to do with the commercial requirements behind it or the raising of the capital, but it's definitely going to solve world peace and hunger and everything else. And look, I would love to be part of that if that can happen. That's why I'm here. But I do demand proof. If you're going to say it, prove it to me. Prove it in our lab, come into the sandbox, show me the data, huh? And then we'll help together. That's kind of where we balance risk.

Speaker B: Just a very quick practical question, like if, uh, people want access to the sandbox, how do they go about that?

Speaker A: Uh, yeah, we have pretty good online publications that explain the services that we have. My job is a little bit evangelistic. I'm going out and trying to spread the word on what we do. Part of it is come and speak to you guys. Because the audience is important to me. People can contact me directly and they often do. So LinkedIn is a good place to find me both in terms of publishing my activity, my thoughts, my focus areas and the work that we do so that you can engage directly there. But obviously the FCA website is a brilliant source of information. Usually what happens is they will contact us through our uh, online service. We will immediately connect up, communicate, explain the services and we then progress through different levels of service. So when we talk about sandboxes, that's a kind of catch all phrase that everyone uses. We do have a regulatory sandbox which we were the first in the world to develop. It has been replicated, as I said at the beginning, by over 90 other regulators. And that's a kind of. I uh, think that's a testament to the effectiveness and quality of the work that we do in the regulatory sandbox. That's one path. There are many other paths, more digital paths, so more prototyping paths. There are labs where we can offer access to our cohorts to work with us. We've even started an agentic academy. So for some participants we've offered them academy training. This is no ordinary training. It's kind of like uh, I guess if Carlsberg made training, this would be the kind of training you would love to have. We bring in all of our friends and partners. So you get cutting edge. Absolutely. That's the minute training in the latest thinking around model deployment, open models, open source software. I mean that's what we love to do. We love to share and to give people the opportunities and the pathways. We even call them innovation pathways. The downside of giving a big service of course is it can be confusing. But what we're doing more and more of is trying to communicate and explain what we do. And the most important thing is don't be afraid to lean in and ask. We are a regulator and some people are a little bit cautious of picking up the phone to a regulator because perhaps that historic view that the regulator could be watching your every move from that point on, recording every conversation and interrogating your back, uh, office. It's more about engagement now. It's more about having the conversations inviting you in. Because actually we need to learn from industry. So that's the non altruistic reason why we do it. It's really so we learn at the same pace as industry.

Speaker C: I just had a quick question to ask and I know we talked about convergence, but if you had to name one of the technologies that you think would have the most transformative impact, which would it be?

Speaker A: Well it's a really difficult question to answer but I think I'm still at the camp that agent technologies. So the shift from AI which you ask to execute to a uh, self executing autonomous technology, one that makes those activities happen on your behalf, that has the potential for um, an immense transformation of all activities across all sectors from defense through to medical and of course financial services. When you're looking at a world which can really open up the access to financial services to all, for me it's an inclusion play the ability to offer highly specialized services which currently are available to very few. When you open those up to the mass market. I think we've always spoken about the ability to do the mass wealth management market. I mean that becomes possible when you can give an AI agent your information and allow it to work on your behalf. Now that's transformation. But the problem with that path ahead of us of course is that's also littered with potential risks. So going back to our last conversation Dave, in order to be able to monitor whether that's risky or not risky, we need to understand it in the lab. So we've just announced our second cohort in the AI lab which is 21 firms looking at agentic commerce, agentic payments, energetic wealth management, all ah, sorts of aspects of this transformational technology from a view where we can work with the data and understand what's happening on the ground level with some of the firms that are deploying that. So I think the difference is we can see the opportunity, the transformation opportunity in that technology. And if there's one, I probably pick that at the moment, maybe next year, the year after it, it maybe shifts to quantum because I think again it's a little bit for this bearer of little brain you get to a point where when you're considering compute that um, acts on that kind of basis, understanding the use cases, even just understanding the experiences. Dave, I don't know about you but it starts to hurt my brain.

Speaker B: Oh yeah. I mean I tried to explain Schrodinger's cat to somebody and then it's like

Speaker A: yeah it's somebody else can my cat ejected when I closed him in the box and tried to explain it minded

Speaker B: when I poisoned it as well. So anyway, no, but I think, I mean it's really interesting you kind of picked on AgentIQ because I think that has the opportunity to be transformational. I think I've written about everything from kind of neurodiversity to basically the fact that the finance industry has not done enough from a consumer customer experience, I think, and this is a big opportunity, but it does open up a world of fraughtness from a right kind of regulatory point of view. So I guess having the sandbox, having these proof of concepts mean that you're able to sort of understand, like if a piece of software makes a decision on behalf of a customer and they don't quite understand what they're agreeing to when they say, yeah, that's absolutely fine, execute, like where the risk in that? I mean, are, uh, you sort of building up frameworks around agentic from a regulatory point of view on the back

Speaker A: of this work, what we're doing? I'm not sure whether framework describes exactly what we're doing, but I think it does because we've got so much practical experience now, because, remember, we started the AI lab 18 months ago and that was quite early days and I don't think there are many regulators operating at that age of technology. In fact, I think we're probably the only one that has a platform like the one we operate. But what it does give us is a really great opportunity to observe and as we observe, we work with firms who are, I'm really proud to say, UK firms are willing to share with us their journey and what we're doing with some of that learning together. It goes back to those first points we raised. We don't do this alone, we do it together. We're going to publish good and poor practice in, I think about four or three months time, that kind of time frame, just in the autumn. And the good and poor practice will allow us to start to share in practical terms what this actually means. Because of course it's going to be focused on the outcomes, it's going to be focused on the principles, but it will have sufficient detail to allow those who are working in this technology to be able to be confident and, uh, comfortable to start to progress some of their deployments into production. We've thought this through. We have the predominant work on the AI lab is in pilot stage, so early stage development, but we have of course AI lab testing as well, which was created specifically to work on that frontier where products are going live and we can look at the customer interaction. So we have got examples where we've started to record and look at how this is actually working in practice. Does it actually deliver, um, for example, those positives like inclusion does it open that window and that door to services that were previously closed to a lot of people. Neurodiversity is really interesting, Dave, because from my perspective as well, we have to be looking at, uh, everyone's access to these tools. So does a defined agentic chatbot give you more access? Does it allow you to have conversations in different ways? And does it give you the education that you need to make good decisions which look after your wellbeing financially? I think also to the point, Dom, of how the regulator can embrace those technologies and use them to police the uh, change. I think another thing we're looking at is a great deal of subtech development, so supervisory technology developments. It's, in a way we're going to have to think about how we deploy that in the future and how we use the same techniques and tools to ensure that there's a safe outcome for the UK system and for consumers. So I'm actually quite positive and quite optimistic that yes, there are risks and there are definitely areas we have to be very vigilant of as we progress. But I think practically working with firms and with technologies, it allows you to be on the pitch playing the same game as everyone else. If you're in the stands and you're looking and you're observing, you're going to be at a disadvantage, I think so. This is why we principles based and we're practical in what we're trying to do.

Speaker C: Just something you said there, Colin. Do you mean when you said about using the technologies for yourself? Would you build something like Anthropic did with Mythos, to detect vulnerabilities in cybersecurity? Would you do the same thing for financial services in terms not of cybersecurity but just of compliance risks, etc.

Speaker A: Well, I mean we've been doing that for a long time.

Speaker C: So that's my automated tools.

Speaker A: As in, well, automated is, I mean more of a question of deployment and choice on deployment. I think it depends on the marketplace and how it's developed. So you can see a situation. I mean, I don't know how deep you are at the moment. On Agentic Payments, for example, there's a lot of work going on in Agentic Payments and we're watching it with great interest. And it's really interesting. If you first think of automated payments, you might think, well, where's the user experience advantage there? It's a tiny bit of the checkout. It's that, uh, last couple of seconds, if that where you press buy. We're not really talking about that. Actually, what we're talking about is what about that new world of micropayments or microservice payments through potentially through stablecoin, for example? What sort of world where your machine is consuming tokens or credits micropayments are going through? There's a world there where actually it would be very difficult to monitor some of that traffic because of the speed and the amount. So there's, of course there are valid experiments in terms of machine to machine monitoring. But I mentioned when I'd worked on high frequency trading in the early part of the 2000s, I remember then people would question whether there was actually any competitive advantage in the notion of a millisecond in terms of high frequency trading. And I think that argument's been made and won and exploited by the trading community. So I think we have to be real about vulnerability and we have to be real about ensuring that we use technology in the right way to help us with our risk monitoring. I mean, I'm working with partners like bis, for example, where we are looking at common standards for financial crime across borders. Wouldn't that be a wonderful world where you could actually stop scams and fraud because you're looking at data patterns and you're able to respond quickly and close down those individuals who are acting not in the interest of the system and trying to scam vulnerable customers, for example. So that's probably why I'm optimistic because I tend to think of technology being ability to be deployed positively in a situation where you might have negative consequences. So I would always err, uh, on that positivity.

Speaker C: Yeah, I was kind of leading to, I guess, can you see a world in the future where the regulators actually have a software agent that is mandated to monitor a bank's other agents that uh, they're running, rather than like sit on the outside, be actually inside the bank monitoring it?

Speaker A: Yeah, I might frame it slightly different to that. I mean, everything is possible with technology. So I've learned a long time ago not to suggest that isn't possible. They're all a possibility. I would more think that if you took an extension of good and poor practice and you created a set of benchmarks, for example, as they do in defence, there are good benchmarks for the deployment of high quality AI in defence of medical, for example, as sectors. Imagine you had those kind of benchmarks within a financial services environment. Then you had an API serving out to firms and you said, look guys, we're looking at 50,000 firms in the perimeter. Here's the benchmarks. Test yourself on These benchmarks and it's an open competition. Imagine a world where you self select the quality of your AI. It's almost like I remember. Do you remember? I don't know whether you remember the days of Mondo. Do you remember Mondo before Monzo?

Speaker C: Yeah, yeah, yeah.

Speaker A: I was working back in those days, I think I was working for Cap and we were doing some work with them and there was a suggestion to publish their backlog to their customers. Do you remember that?

Speaker B: Yeah, yeah, yeah.

Speaker A: Hugely successful, brilliantly successful. Right. And it was so, uh, I mean, honestly, I was working I think at the same time with an incumbent who almost fell off his chair. When I showed it to, he was like, how could you possibly do that? Giving your customers access to your potential future features and your competitors. Anyway, as you say, Dave, uh, it was hugely successful. Built quite a element of trust in the product and a lot of loyalty. So I look at the system and I think, wouldn't that be wonderful if you could have a system which is in a way self regulating rather than being a bit. The conception of agents going out and chasing down rogue financial influencers, for example. It's a little bit Terminator esque for me, Dom, although I am a fan of Arnie, but I think there is something to be said with applying robust technologies to create a world where self regulating systems actually become the norm.

Speaker C: Yeah, yeah, that's very much where I was coming from. Like in a, uh, technology sense we might have self healing systems that repair themselves but my view was like, well, we should extend that to regulations. So.

Speaker B: But it is interesting because let's. I don't know if you remember, Colin, I wrote this article, the Three Body Problem, which was if you've got a consumer agent and a business agent, they're interacting. The consumer agent wants to feel. The business agent wants to extract the most value out of that. That happens like on behalf of the business and the consumer. I did imagine actually a third one which is the regulator making sure that they're playing nicely. Because whatever we do, we're going to start building our biases as sort of consumers or from a consumer point of view or a business point of view into the agents and what they represent. So whilst I can understand the sort of notion of self regulation, I also think actually having an umpire may be something that's kind of a necessity as well.

Speaker A: Agreed, agreed. I was watching the, the very, very breaking news from Yorkshire and the cricket umpire who's now being questioned because the guy behind clicked his fingers. I thought that was a lovely Analogy as you were speaking about. So did you see this thing on uh, well, it's in the second division of cricket and someone was called out, but it was looked at the video as though one of the fielders clicked his finger at the same time as it clipping the bat potentially. So if perhaps the umpire picked up a, maybe a rogue click. What an amazing analogy for what we're talking about. How do you get an umpire that's sufficiently good of hearing that he can distinguish between a click and a bat that's been. Then he makes the judgment correctly. Maybe we need var for a, uh, cricket or something, I don't know. But I love the thought experiment. Dave and Dharmish, I, and I appreciate you giving to me. I have also played with the thought experiment around decentralized regulation as well. Decentralized autonomous regulator, a dar, not a Dalek, a dar. So I think there's so much we can think of in future iterations of technology. I think the only thing we can do as a tech positive forward thinking regulator, uh, we talk a lot about being a smarter regulator. And for us that strategy of being smarter, that word is quite critical. So whether you do it through autonomous or through deployment of more standard technologies or deterministic technologies where you can use quantum to look at vast amounts of information and predict future trends, because there is something of the Minority Report that you could bring into this discussion where you could start to predict patterns ahead of time, some kind of patterns which, I mean I've certainly looked at superhuman prediction modeling and there's some incredibly interesting technologies coming in that area. So I think we could probably spend another hour talking about that.

Speaker B: But yes, please, at some point that would be, I think it's fascinating. I mean I feel like I've learned a lot about, I guess, the role of the regulator. I always love going to other places around the world and when we have meetings to be told that we've got the best regulation in the world and models, our model around open banking is one that others are adopting. So I think many congratulations to the team. I guess it's sort of great to hear that we're being used as a model around the world. It's great to hear that you're sort of, you see your role is not just about the regulation. It's about helping companies understand a bit about the future and the role of the consumer and risk and all of these things in the future. So I think that's been really, really good to understand. I had one question which is there anything about AI which You think the banks just aren't getting and it's frustrating to you. Is there anything in particular that you just sort of think oh come on guys, you really need to get.

Speaker A: I mean, I haven't seen that evidence and I suppose it's not really for us to judge that on their behalf. I think from my perspective, those that we are engaging with really do have a positive attitude to the technology. I think partly because banks and financial services have used AI for uh, 30 plus years. So I know that was deterministic AI and now we're in a world of probability and probabilistic activity, which is different. I think we all accept that there's a shift, which we've been on the journey of for three or four years. But I honestly say, number one, I don't think we would judge whether people don't get something or not. But there may be businesses that choose to have a competitive advantage by not deploying certain technologies. You've had examples where certain brands have gone against the grain when it comes to experiential activity. You might look at a Metro bank is a really good example of some brand that went into branches where others were uh, rolling out of branches. So there are some times when the digital nirvana, and actually this is something Dom and I have often spoken about in a way, the frothiness of new entrants, of course, the work that goes into promoting them. It's terribly exciting, it really is. And it attracts the headlines and others really working diligently to uh, maintain really good relationships with their customers, deploy really good technology. And I think of like the building societies, those credit unions who have amazing opportunity. The mutuals is an unrecognizable sector. Uh, if you want to compare it with how it's developing. You might have thought five years ago it was a laggard. Now it's taking great advantage of its position within the community, that trusted position. And I think that's got a great advantage that we're going to really work with. So you might have said a few years ago, well, there are some areas of the economy which perhaps haven't focused on it enough, that maybe they haven't deployed the technologies fast enough. I think all of these things, they have their pros and their cons. I personally think that our job is really just to demonstrate, to offer the services that others can come and consume which will give them the support to do what they can do. And the biggest thing is for me to communicate that as widely as possible so that the industry hears directly that the regulator's there to work with them and to give them that boundary within which they can experiment safely and then they can deploy these technologies in a responsible way. That's my job. And if that brings companies with me and it helps them, I will have done my job well and I will be absolutely be able to sit back and say, okay, well, the UK is doing well when it comes to it. As you say, Dave, the perspective from others around the world, we are punching above our weight and we are really focused on the outcome for the consumers of the UK and the wider system. So, yeah, I think there's nothing I'd say beyond that. I think the firms we work with, I've been pretty impressed by the engagement, so I can ask no more than that. When firms come to us, they're open, they're engaged, they're willing to share their experiences. A regulator can't ask for much more.

Speaker C: Fantastic.

Speaker B: Thank you so much for your time. That's been an, uh, awesome conversation and there's probably gaps, and I'm sure there'll be sort of potential. Well, I'm hoping there'll be potential to have another conversation in the future, but, you know, it's really good to kind of hear what's been going on.

Speaker C: Really appreciate your time, Colin. Thank you.

Speaker A: Uh, it's a pleasure. It's great speaking to you. Uh, always love you to catch up and hopefully see you in the flesh quite soon as well. Thank you for tuning in today, Dave and Dharm M. Demystify. We hope you enjoyed the show. Don't forget to like and subscribe and tune in next time as we take another topic and demystify it.