One of artificial intelligence’s key strengths is its ability to spot anomalies—a functionality that Bank of England Governor Andrew Bailey said banking regulators aren’t fully leveraging.
Bailey called for greater investment in AI and data analysis, despite the substantial investment many central banks have already made in the technology.
The regulator noted that in many cases, current models are generating vast amounts of data for regulators to sift through, but that, “none of us, I think, can put our hand on our heart to say that we’re sort of optimally using it all.”
This inefficient analysis of data, even with AI, raises concerns that there could be a “smoking gun” right under authorities’ noses—such as evidence of fraud or money laundering in the financial institutions they are tasked with overseeing—that they are unable to pinpoint.
Evident Fraud Protections
The significant benefits of deploying AI in fraud detection have become more evident as the technology sees wider adoption.
According to a FIS survey of business and tech leaders, over three-quarters of respondents said that AI enhanced their organization’s fraud detection and risk management programs. As a result, nearly half of these leaders indicated that their companies plan to increase AI investment over the next two years.
A separate study from the Bank for International Settlements (BIS) and the Bank of England found that AI models are a valuable fraud detection tool, even when analyzing real-time payments. AI not only proved more effective at detecting suspicious activity than traditional fraud defenses but also enabled financial institutions to uncover new fraud patterns much faster.
Actively Addressing the Issue
Although AI has been a game changer for fraud detection, it has also been a powerful tool for fraud perpetration.
Bad actors have been able to adopt AI much faster and at a larger scale than the financial services industry, as they are not constrained by compliance or regulatory requirements.
Both financial institutions and their regulators have often been overwhelmed by the volume of data AI can generate and unsure how to process this information or integrate it into their day-to-day operations.
Many banks and credit unions have also been hesitant to give AI free rein in fraud detection due to concerns that the tech could produce false positives, which may increase customer friction.
However, the growing threat of fraud suggests that consumers may be willing to tolerate occasional false alerts in exchange for stronger protections. According to data from the University of Notre Dame, most consumers stay with their bank if the institution actively supports and protects fraud victims.
