Digital Banking

Turning To AI To Reduce Digital Banking Risks

Implementing AI might be all the rage, but are banks buying too much into the hype? FIs need to cut through the noise and find use cases which solve their biggest pain points, says David Berglund, SVP and AI leader at U.S. Bank. In the latest Digital Banking Tracker, Berglund explains why a “Skynet of robots” powered by AI is closer to science fiction than reality, and how FIs can better use AI for use cases like fraud prevention and risk reduction.

The financial services industry is booming. It brought in $145 billion in 2017 alone, with revenues projected to continue to increase in the coming years.

Risk looms large with rampant growth in fraud, however, and financial institutions (FIs) must stay guarded to prevent cybercrime. Banks of all sizes are now looking to reduce the risks they face when doing business, increasingly turning to new and emerging technologies like artificial intelligence (AI) and machine learning (ML) to reduce risk and combat fraudsters.

According to a survey conducted last year by the Global Association of Risk Professionals (GARP), 88 percent of bank executives believe AI and ML adoption could provide “a foundational change” for risk management. Both enable banks to dig out suspicious transactions, something many other solutions could not, according to David Berglund, senior vice president and artificial intelligence leader at U.S. Bank. In a recent interview with PYMNTS, he noted the development could help FIs find and recognize common elements in legitimate payment data.

“AI is really good at finding patterns in large sets of data and lowering the costs of making predictions [based on that data],” Berglund said. “The question becomes how you use those strengths and capabilities, and that’s really up to each individual bank.”

Reducing Risk By Fighting Fraud

One of the largest, still-growing risks that banks face is increased cybercriminal attention and attacks. FIs stopped nearly $17 billion in fraud activity in 2016 — the most recent year that statistics were available — and lost $2.2 billion to it during the same year. Both were notable increases from 2014, which saw $11 billion and $1.9 billion, respectively.

Banks and solution providers can use AI- and ML-based systems to identify elements of a transaction that typically indicate fraudulent activity, such as a transaction conducted using a newly created account or obscure email address, Berglund said.

“We’ve got several different advanced fraud-fighting models that are looking for out-of-pattern spending, or spending that happens in certain geographies and areas where fraud seems to be higher,” he explained. “[They] can figure out when something seems suspicious and flag it.”

These models can also evolve over time, meaning AI- and ML-based systems can map out recurring or shared transaction elements to build new guides and rules for flagging suspicious transactions. This ability to ingest new information allows FIs to create automated security processes that can be performed without forcing customers to endure long, complicated authentication checks, making them more likely to accept advanced security offerings.

“That helps us soak up new data points, and helps us further refine that model and apply those data points to improve our customer experience,” Berglund added. “For example, we’ve learned that if customers share their mobile locations, we can compare card transaction data to [those] mobile locations and the model will know if the customer is actually making that transaction. [It will also know] if they’re a thousand miles away, without the customer having to do anything else.”

AI’s Long-Term Impact

AI and ML may be the latest tools in the fight to reduce fraud, but they are far from fully evolved — and could never fully replace human expertise, he noted.

“A lot of people nowadays think of AI as Skynet robots that are going to replace or take over humans,” Berglund said. “But, that’s really more science fiction and a pretty extreme view. That probably isn’t going to happen anytime soon, or ever.”

That’s not to say that AI can’t change how banks secure transactions and reduce risk, though. He predicted that AI and ML technology would, most notably, impact the space as tools that human employees and experts can use to expedite credit reviews and give a more accurate view of a potential client’s creditworthiness, among other potential risk reduction use cases.

“A lot of times, AI gets viewed with a lot of hype,” Berglund explained. “But, as a bank, we’re really thinking about deep learning: how computers, machines and devices can learn from data and come up with conclusions that we can continue to refine and improve as these machines soak up more data.”

While the FI-inspired version of Skynet isn’t likely to debut anytime soon, AI and ML are already making their presence felt in the banking space.

First Name*

Last Name*

Title*

Company*

Work Email*

BY COMPLETING THIS FORM, I HAVE READ AND ACKNOWLEDGED THE TERMS AND CONDITIONS .

. . . . . . . . . . . . . . . . 

About The Tracker

The Digital Banking Tracker™, powered by Feedzai, brings the latest news, research and expert commentary from the FinTech and consumer banking space, along with rankings of over 300 companies serving or powering the digital banking sector.

——————————–

Latest Insights:

Our data and analytics team has developed a number of creative methodologies and frameworks that measure and benchmark the innovation that’s reshaping the payments and commerce ecosystem. The PYMNTS Next-Gen AP Automation Tracker, is a monthly report that highlights the most recent accounts payable developments and automated solutions that are disrupting how businesses process invoices, track spending and earn rebates on transactions.

TRENDING RIGHT NOW

To Top