10 September 2024

Understanding Artificial Intelligence in Financial Services Research

This report looks at current and future use of artificial intelligence (AI) within the financial services sector.  

The report draws on responses from 13 finance representatives in asset management, banking, financial advice, and insurance. Their consensus was clear: AI is a pivotal technology that holds significant promise. All participants either have integrated generative AI into their operations or plan to do so soon. They report being driven by motivations such as enhanced customer outcomes, improved operational efficiency, and sophisticated fraud detection.

We hope this research is a helpful conversation starter. While this research is not guidance for financial service providers, we are committed to working with the industry to ensure that technological innovation delivers positive outcomes for New Zealanders. 

Download FMA Research Publication: Understanding Artificial Intelligence in Financial Services [PDF] 1.6MB

Jess

Kia ora, I'm Jess. Welcome to our second podcast in our Occasional Conversation series. I'm here today with Stuart. Stuart, can you introduce yourself to our listeners, tell them a little bit about your career and your position at the FMA.

Stuart

Thanks very much, Jess. My name's Stuart. I'm the Chief Economist at the FMA. I've been at the FMA about 18 months. I run a research and economics team and a data science team. Together, these teams investigate and undertake research into the financial services sector with the overall aim of improving outcomes for customers. Before I joined the FMA, I used to live in London. I worked for HSBC. I was the head of behavioural economics for a number of years. And I also worked for Aviva as their head of conduct and the head of the customer engagement engine, which was a a big piece of AI that used machine learning to identify customers for different products.

Jess

Can you talk a little bit about the research that we're going to dig deeper into today?

Stuart

For sure. So, my Economics and Research team did a piece of research entitled Understanding Artificial Intelligence in Financial Services. Really, the research aims to understand how banks, insurance companies, asset managers and financial advice providers are using AI in their day-to-day operations. And as you would expect for a regulator, we're super interested in how these firms are managing the risk around AI, so the research focuses quite heavily on risk management.

Jess

And why is this AI research so important to the FMA?

Stuart

Yes, it's a good question. So really for three very specific reasons, we wanted to understand how the industry is using AI. So we had some great stories about how the industry is using AI in chat bots, how they're using it to draft the first version of customer communications. We wanted to understand the risk management and we heard some good stories about how firms are using their current risk frameworks and their current risk environment to do that. And perhaps most of all, we want to understand and ensure that New Zealanders have access to the very best technology, and so the products and the services that are being delivered by the industry are as good in New Zealand as they are anywhere in the world.

Jess

And I'm interested to know a little bit more about what the key findings were in the research.

Stuart

Ah, yes, Jess. Everyone's always keen on the key findings. So something that I was really pleased to discover is that nine out of the ten firms who are using AI right now are using that AI for improved customer outcomes. And so that's a situation where the firm can apply AI to deliver a better outcome for a customer. So examples of that, as I said before, are things like chat bots where the AI can undertake very, very simple processes, often through an online chat bot to help customers in their banking or their insurance needs. If we dig into the findings a little bit more, we spoke to firms about how they're currently using AI and many firms touched on the importance of fraud detection. One of the things AI is really good at is looking for patterns and identifying patterns in data, and it turns out that's really useful in fraud detection. Another area that many firms touched on was around risk management and using AI to identify new risks. Or perhaps the articulation of risks across their business. Finally, just let me touch on some of the future activities. All of the firms that we surveyed or spoke to talked about using AI for their customer interactions in the future. And many of those firms also touched on operational efficiency and you know, operational efficiency is something that economists, they're always going on about talking about efficiency and markets efficiency in firms. Well, what does that really mean? Efficiency really means doing something better, or, as the case is doing something cheaper. So when a firm is using AI to cut down their costs, well, that's great for the shareholder because probably we're going to see their dividends go up. But it should also be great for the customer, because hopefully fees and charges on the products and services that the firm is providing should fall. One of the things I'll be really interested in moving forward is to think about how AI is delivering cost savings and then ultimately decreases in fees and charges. Right across New Zealand's financial services industry.

Jess

You talked a little bit about risk there. Would you be able to go into a little bit more detail about that and what those key risks were?

Stuart

One of the challenges with AI is that we internationally really are at the beginning of the journey with AI. When a firm starts applying that technology across its operations the risk environment changes and we're at the beginning of understanding the risks that AI poses and how we can manage those risks. Jess, you asked about the broader risks. And it's all very well and good for me to talk about the current risk and compliance control environment, but as AI develops, we run the risk of getting into this idea of the black box where AI becomes so complicated that we don't understand what's going on. And that changes the risks. One of the things that we've been talking to lots of firms about is how they manage that risk when they buy AI into their organisation and if we think about IT, firms traditionally buy in all sorts of different skills technology and they need to manage the risk there. Firms have got reasonably good outsource or third party risk management systems and some of our early conversations with firms have been about how those systems apply to AI vendors just like they apply to traditional technology vendors. It's a different part of risk management, but it's a really important part in the context of buying in AI services into firms.

Jess

I find that just so interesting. There’s is just so much we don't know about AI, so coming back to the research, what do we want New Zealanders to know or what do we want for New Zealanders?

Stuart

The most important thing for New Zealanders is that we want New Zealanders to have access to the best technology and to products and services that leverage that technology in New Zealand. Lots of providers are doing as much as they can to bring in the best technological solutions, and that's fantastic. But I have to be honest, I get a little bit frustrated when I see products and services overseas that aren't available to New Zealanders and so we need to adopt AI in a, in a way that is safe, in a way that takes account of the risks, but we also need to remember the efforts. And so the most important thing, Jess, is that New Zealanders have access to the best technology within financial services.

Jess

Yeah, totally. And is the FMA going to publish some guidance around AI do you know?

Stuart

Ah yes, guidance well. AI is an amazing new technology, but as I said, we're at the beginning and so I think it's a little bit early for the FMA to be thinking about guidance. I think it's really good that we're talking to industry and many of the people in the industry that I've spoken to are really happy and really welcome the conversations. I'm looking forward to working closely with industry, but I think it's far too early in the development of AI for us to be thinking about guidance. Maybe one day on the horizon, but I doubt one day soon. Jess

Thank you so much for your time today, Stuart.

Stuart

Thanks Jess.