Algorithm-based overdraft protection lacks disclosure and transparency

There’s no doubt that algorithms can reach levels of analysis that humans alone can’t make. In banking, Artificial Intelligence (AI) might identify fraudulent activity or improve fair lending practices — both of which ultimately offer benefits. Algorithm-based models have their limitations, because they often have murky calculations when it comes to overdraft protection. Plus, they can damage your account holder relationships and hurt them financially.

Last October, CFPB Director Rohit Chopra said that the rapidly changing payments landscape, notably Google and Apple Pay, is changing the way consumers engage with money, which presents challenges, especially for smaller financial institutions. The challenges have contributed to a shift from relationship banking toward algorithm banking, which can have detrimental effects on consumers.

The negative impact of a shift from relationship banking toward algorithm banking and AI-driven models

Undoubtedly, the payments landscape has changed dramatically. Accessibility and convenience have pushed digital wallet purchases to more than 48 percent of global e-commerce transactions.  According to The Global Payments Report by FIS, global e-commerce market is projected to grow 55.3 percent between 2021 and 2025 to reach about $8.3 trillion in transaction value.

Financial institutions turn to AI-driven models of overdraft protection most frequently when implementing a dynamic limit strategy. Rather than applying a fixed overdraft fee structure, an algorithm uses variables such as overdraft history and consumer behavior to change the limits of the overdraft. With consumers relying heavily on digital payments for online and in-store purchases, it can be challenging to understand their account balances. And it is even harder if they don’t understand their financial institution’s overdraft program.

 

continue reading »