Podcast: Play in new window | Download
The year 2022 was one of global financial uncertainty and risk, and 2023 may bring more of the same. For executives in payments risk management, planning for the year ahead should involve taking this geopolitical and financial uncertainty into account and planning accordingly.
In a recent PaymentsJournal podcast, Sudhir Jha, senior vice president and head of Brighterion, a Mastercard company, and Brian Riley, head of credit and co-head of Payments at Mercator Advisory Group, discussed how artificial intelligence (AI) is growing as a tool in payments risk management, and they also delved into the key trends they expect to see this year.
Security Challenges Facing Financial Institutions in 2023
There is likely to be a heightened risk of geopolitical conflict, inflation, and resulting credit issues in 2023. “Unemployment has been low, but household budgets are not keeping up with the cost of living,” Riley said. “Interest rates are off the charts.”
In the past year, there has been an increase in scams associated with peer-to-peer (P2P) payment apps, such as Zelle, a trend that’s likely to continue.
As a result, many are looking to AI to help prevent such scams. “The developments in AI are not accessible to everybody,” Jha said. “The government hasn’t done the equal investment to make it available to everybody. In the ‘70s, ‘80s, and ‘90s, there was a lot of funding of fundamental research that could be made available to everybody to use. From a government perspective, the U.S. is not investing enough in AI to make sure that the research trickles down to everyone.”
Smaller players don’t necessarily have the capacity to develop their own AI solutions. As a result, they often form partnerships with larger companies, such as Mastercard, to make use of the larger firms’ increased computing power.
Brighterion is using AI to address P2P fraud across its network. “We are building solutions very similar to what we have done for the card side, on the account-to-account side, to provide solutions that work at the network level,” Jha said. “We can give you a solution that works across card payments, ATM transactions, and even crypto transactions. Fraudsters are aware that many financial institutions have silo applications for finding fraud in one channel, so they will have one card for account-to-account and one for crypto.
“We are trying to provide a multichannel solution for both fraud and money laundering to provide full visibility for all transactions, and helping to capture fraud across the network.”
Fraud Scenarios We Expect to See
As AI-powered fraud detection software has gotten more sophisticated, fraudsters have become less successful at creating synthetic identities that pass detection and have moved more toward scams. “They’re trying to make the person who actually owns the instrument—whether that’s a cellphone or credit card—do something that is not in their best interest,” Jha said. “For example, if I were a fraudster, I could call you and tell you that you won $10,000, but for me to give you that prize money, you have to send me $200 for shipping.”
New machine-learning algorithms are being developed to give people warnings when they attempt to make a transaction that appears suspicious. “We have to figure out a way to flag it for them, and change their mind,” Jha said.
According to Riley, faster payments and P2P payments can make scams even trickier to combat, because payments can go directly between bank accounts instantaneously. “The payment transaction, unlike a credit card, is irrevocable,” he said. “To undo that whole mess takes a lot of work.”
In the coming year, scams on P2P platforms will accelerate because they are still very hard to catch. “There will continue to be payments risk in the entire market, whether it’s at the merchant level or the consumer level, and there will be a range of issues due to the impending economic slowdown,” Jha said. “And people already have spent some of the savings that they had. That creates a credit delinquency issue, which leads to even more fraudulent claims by the merchant or by the consumer, and more openings for scams. The only thing we can really do is suspect that this is not a transaction that you normally do and warn you that it could be a fraudulent transaction.”
As AI develops, it will get better at detecting scams, but that development remains in a nascent stage. Financial institutions can look forward to better tools in the future, as AI solutions spread throughout the economy and use cases multiply.
