As the nascent artificial-intelligence field develops, financial institutions are readying to adopt AI in their fraud-prevention measures just as criminals are devising AI-enabled ways to make their schemes more harmful. Evidence of this is that 43% of financial institutions surveyed for the DataVisor “2025 Fraud and AML Executive Report” will deploy generative AI into their fraud-prevention services in the next year or two.
Though none say they have fully implemented generative AI tech, 33% says they have done so in some areas, with 6.2% in the implementation phase. Only 21% indicate they have no plans to use generative AI in this capacity.
“AI has become a double-edged sword in fraud prevention,” Yinglian Xie, chief executive and cofounder of DataVisor, says in a statement. “Fraudsters are innovating without regulation or legacy constraints, while organizations are still working to scale AI defensively.” DataVisor is a Mountain View, Calif.-based fraud- and risk-management platform.

How much of an issue is criminal use of generative AI technology? A lot, apparently, with 43.8% of respondents saying the tech benefits fraudsters and criminals compared with 52.2% saying they are either unsure or that it benefits both criminals and organizations equally.
Eventually, however, financial institutions expect a more balanced benefit with just 12.5% expecting criminals to benefit in the future and 81.3% expecting the benefits to be balanced or are unsure. Just 6.2% says organizations will have the upper hand.
One expectation for how generative AI could help banks and credit unions is with lowering manual investigation efforts, which most—68.8%—cited in the report followed by enhancing detection accuracy, 56.3%, and reducing false positives, 56.2%.
The DataVisor report also found that first-party fraud is a major challenge for financial institutions, with 68.8% declaring the increase in first-party fraud as a top challenge.
The increase in U.S. credit card balances is one factor in why banks and credit unions should pay more attention. With U.S. households carrying $1.18 trillion in credit card balances, according to the Center for Microeconomic Data at the New York Federal Reserve, consumers with higher balances could resort to first-party fraud should the U.S. economy incur a downturn, DataVisor says.
Another factor for why this fraud is important is that regulatory shifts in authorized push-payment fraud liabilities could fall on banks and payment processors, the report says.

