AI and the Future of Personal Finance

Artificial intelligence seems to be on everyone's minds. It's increasingly common across healthcare, education and business, where it's used to rapidly process and analyze large data sets, simplify repetitive tasks and provide guidance tailored to the user.

The personal finance world is no exception. AI is already used to help detect fraud, run chatbots and help with underwriting decisions at many financial institutions.

Nadim Homsany recently joined BECU as senior vice president of AI strategy and innovation and is the former CEO and co-founder of EarnUp. He and his team are developing a generative AI tool for BECU.

We sat down with Homsany to talk about how this emerging technology could help improve financial health by increasing access to financial systems, products and services and discuss how the regulatory landscape might evolve to ensure people's money is well protected.

Q: What personal finance problems do you think generative AI has the potential to help solve for people in their daily lives? What is the need?

A: The broad answer is that AI can increase access to personalized financial products and services for more people to support them in improving their financial health. It can increase efficiency and streamline processes for employees, and that reduces costs.

I think the more specific, narrower answer related to financial health is that if we can get people the right financial guidance when they need it, then they can make better financial decisions for themselves to improve their financial well-being.

A lot of people think the silver bullet for financial health is financial education, but studies show that financial education by itself doesn't work. Where it does work is at the point of decision making. So, if you can get the right financial guidance at the right moment in time, people make really, really good decisions.

That can be hard to do when you don't have the wealth or time to get that guidance from a human expert, or you live in a financial "desert," where you have to drive for 30 minutes just to go see an advisor.

We all live increasingly fast-paced lives. Even if you do have a human financial professional, you may not have time to call or email them. If you're in a hurry and just trying to check off items from your to-do list, you're less likely to make the best decisions.

Q: What are the benefits of AI for financial institutions?

A: Financial services exist in a complex regulatory environment — by necessity, it's one of the most highly regulated industries in the United States. Because of that, providing any services in a low-cost manner is very hard. AI can bring a lot of efficiency to the table and reduce the cost of providing those services.

Last year, we provided more than $490 million, in total, of benefits to BECU members (through great rates and low fees). If there are ways to increase efficiency in the industry so we can provide greater benefits at a lower cost, that's all good.

The real advantages of AI for financial institutions are in the improved benefits we can provide to our members. We can be available 24/7 to provide personalized service without adding operational cost. We can also automate and speed up routine tasks and compliance reviews.

Financial institutions that use AI have a potential competitive advantage — differentiating their products and services with faster product development, better fraud detection, audit automation and regulatory compliance.

Q: Money can be an emotional topic. How do you address people's concerns about using this new technology to manage their money?

A: The best way I have found over time to alleviate people's concerns is just being very transparent with them as to what's happening with their information. I think we'll find that — in at least the medium to long-term — we're not going to be doing anything that differently here than what has already been done with data.

Q: How are you thinking through issues like responsibility for potential mistakes an AI tool might make and any negative financial consequences for the user? What can financial institutions do to mitigate and respond to those risks?

A: Through testing and development, we're looking at how to increase accuracy. AI is likely going to help us process existing regulations and adapt to new regulations faster than we otherwise would be able to without the technology.

An example would be how to review marketing materials for compliance with Unfair, Deceptive, or Abusive Acts or Practices regulations. In many businesses, every piece of marketing has to go through UDAAP review by a compliance officer. That takes time, and for many small businesses, you need to rapidly iterate on your marketing materials — so that can slow down the whole business.

An AI system might be able to provide a UDAAP judgement in a second or two. Think about how much more effective and efficient that would be, and how much happier people would be in their jobs and the regulator would be knowing that there's a tool out there helping them do that.

In terms of negative financial consequences, I think at first, while the deployment of this technology in the broader financial services ecosystem is still new, we'll rely on existing regulations to respond to any financial consequences to the user. That will probably change over time as the use of generative AI becomes more common and people become more comfortable with it.

Q: Beyond the existing regulations for financial institutions, do you think we need specific regulations for AI?

A: There absolutely needs to be regulation of AI. I think the critical piece of regulating AI is to ensure that the regulation is clear and that it is equally applied across the industry so there isn't any gaming of the system. We can't let AI run free and at the same time we can't impose so much regulation that it stifles innovation.

A useful analogy is the evolution of the auto industry. Since the Model T, cars have transformed modern life — it's hard to imagine a world without them. Safety regulations didn't stifle innovation; they helped ensure technology developed in ways that benefited everyone. The same principle applies to AI: Smart regulation can guide progress without letting it spiral out of control.

The more these AI models advance, the more powerful they become — which is exciting — but the more important regulation will be.

Q: Can you talk a little about what you're building for BECU?

Q: How do you 'train" the AI?

A: For BECU, it will be a combination of in-house and off-the-shelf data and technology. We'll use readily available technology to teach the AI how to think, how to speak, how to structure thoughts as if it is a person with a very good general education. Then we'll use our own technology, engineering and data to take the next several steps.

It's kind of like starting off with an associate degree or bachelor's degree, and then training it with a master's in financial guidance.

In terms of data privacy, consumer data will still be kept confidential. We currently use a bunch of data to inform our machine learning models and underwriting models in the business. It's all anonymized and can't be traced back to an individual. It's like the way we review and grade calls for training purposes. If anything, we'll be doing better at it than the existing technology we already use, partially because it will be updated technology — these large language models contain so much information and are able to learn so fast.

Q: What can we expect from an AI-driven financial tool in the future?

A: There's push and there's pull. We already discussed what I am labeling as "pull." For example, you need to change the tires on your car and you're unsure if you can afford it. Many Americans can't afford a $400 unexpected expense without getting a loan or asking family for money. So, the pull would be seeking that human-like support when you need it and on your own terms.

And then there's "push." Here, we're envisioning a mobile enabled system that pops up on your phone or your smart watch when something is happening with your finances — like your credit score. Or you could choose to enable geolocation, so when you walk into a coffee shop, you can set it to let you know that you only have $5 budgeted to spend on coffee, or you might ask for a reminder every time you walk into a coffee shop to decide if you should buy a coffee that day.

It could also be used to help with savings, so, if you pay $3.75 for a coffee, it can ask you to round up $0.25 and send it to your savings account or to your highest interest loan.

Q: What excites you about the use of generative AI in personal finance?

A: The cool thing about AI and large language models and the way we plan to use them is that you're democratizing financial access. People who couldn't afford financial help now have access to it. There's that old saying: "It's expensive to be poor." If we can find a way to make that access hyper personalized, which is what we're building for BECU members, then that changes the game for people.

Related Content