The Finance Sector Is Hitting an Inflection Point With AI

If you are among the majority of Americans who use mobile banking services, you’re probably familiar with the facial or fingerprint recognition tools that let you access your accounts with a single look or touch of the screen.

A national survey conducted last year by Morning Consult and the American Bankers Association found digital banking to be the most popular method among U.S. consumers, with only 8 percent of Americans visiting in-person branch locations.

MSU Federal Credit Union, which of course offers online banking, also has an AI feature on its app that tells members if there is a better way to shop at their favorite grocery store, offering a guide on which cards to use with which vendors. The credit union is hoping to ramp up those capabilities in the near future.

For example, members may soon be able to avoid going to branch locations to facilitate large wire transactions or verify your identity with a wet signature altogether. Or at least, that’s what Benjamin Maxim, the chief digital strategy and innovation officer at MSUFCU, is hoping members of the credit union will be able to do thanks to artificial intelligence.

“If we’re calling to verify who this person is, we could have them simply do a biometric face scan that matches up with what we’d have,” Maxim told Newsweek. “It would be way more efficient and effective than asking 20 questions about your account.”

Using biometrics to bypass routine identity verification would be yet another step toward streamlining the banking process so that customers and agents can have “more meaningful and longer conversations.”

AI has touched nearly every cornerstone of American society, but it’s long been a natural fit for the finance industry, which is data-driven and typically involves analyzing large swaths of information, identifying patters and making predictions. Banks and investment firms have automated credit scores and fraud detection since the mid-1980s, and those efforts only grew with machine learning in the early aughts, and the rise of AI-powered FinTech in the 2010s.

“AI has been used for a very long time, particularly when it comes to financial modeling, risk modeling, underwriting, that kind of stuff,” Lauren Clement, the vice president of emerging technology and innovation of Prudential Financial, told Newsweek. “There’s a very long history of using machine learning and data science for portfolio assessments or underwriting projections.”

As other industries figure out how to begin implementing AI more broadly, the finance sector, with its deep familiarity with AI, is in a unique position where it’s trying to figure out what happens after AI becomes the norm.

Where Does AI Go From Here?

The industry, Clement said, is at an inflection point: Will emerging technologies, like generative AI and large language models (LLMs), be added to existing AI use in increments, or will they have the potential to leapfrog those more traditional uses and create an entirely new world of AI in finance?

“How smart can these models get?” she asked.

The real-life financial advisers working at insurance companies, like Prudential, currently have tools that allow them to use AI to run enhanced searches, making their work both faster and more detailed. But these institutions are still relying on human advisers to serve as the “main deciders” on which search results to use. Over time, however, the question will become whether there will be an automated model that is reliable enough to make those decisions or if “the human will always be in the mix,” Clement said.

Despite all the advancements, Clement acknowledged that the industry faces three major hurdles in implementing AI. Because financial services have been digitized far longer than any other industry, a lot of older documents are often siloed, posing barriers to adopt tools that need easily accessible and clean data that’s all stored together. Regulations are also currently unclear and vary drastically from location to location, leaving many guessing about the right ways to introduce new AI capabilities. On top of all that, industry leaders need to find ways to explain what exactly is being implemented and just how reliable it is before any of these tools are going to be put into production.

Apart from the logistics of implementation, there are a lot of institutions that seem to have an aversion to embrace AI over data security concerns, Maxim said. In response to those fears, MSUFCU created an AI governance committee that included representatives from risk, legal, human resources, innovation and digital experience.

“It’s just a new way of looking at these things that we’re thinking about and really focusing on acceptable use of AI,” he said.

Will We Still Talk About AI in 10 Years?

While a lot of the conversation around AI feels like new terrain, it’s likely that one day there won’t even really be a distinct discussion about it.

After all, there was a time when you could only find a few Americans online. In 1983, just 1.4 percent of U.S. adults used the internet, polling firm Louis Harris & Associates found in the first survey of its kind. By 2014, 87 percent of Americans described themselves as internet users, according to the Pew Research Center.

“It’s kind of like when the iPhone came out. Everyone was scared of it. Now, all our members and big customers, they walk around with all their financial data on their phone,” Maxim said. “What happened? Well, now, we wouldn’t think anything of that. Now, we think it’s great, it’s easy. In the next five to 10 years, AI is going to be that.”

On the other side of the coin, there are financial companies who inflate their AI adoption, attaching the word onto any new advancements, even if AI isn’t being used to make processes smarter or more rigorous.

“A lot of people claim to be AI. They maybe don’t fully understand what it is, and they use it as a buzzword,” Maxim said. “‘AI’ is the word they’re adding, but really, they’re doing something else.”

That’s why Maxim will be looking for companies that are using AI to meaningfully affect the outcome of their solutions when he will serve as a judge for Newsweek’s AI Impact Awards.

The finalists and winners of the awards, which recognize unique and innovative AI solutions that solve critical issues or advance capabilities across various industries, will be announced in late May ahead of the AI Impact Summit in June. Clement will also sit on the panel of expert judges led by Newsweek Contributing Editor Marcus Weldon, an AI scientist and former president of Bell Labs.

Entries are open until April 25. For more information on the event and entry guidelines, please visit the AI Impact Awards homepage.