Financial institutions are cautiously deploying AI to crack down on fraud associated with digital assets. 

Crypto-related scams, ransomware, darknet markets transactions and money laundering cost financial institutions $154 billion in 2025, a 162% increase from 2024, according to blockchain company Chainanalysis’ Jan. 8 report.

Banks are gearing up their infrastructure to help customers transact, store and invest in digital assets, which includes developing better fraud and anti-money laundering processes, Scott Southall, managing director at Citi Services, told FinAi New. 

“We’ve seen AI tools being used in the industry for things like fraud monitoring and transaction monitoring for some time,” Southall said. “They’re fundamentally more efficient at identifying meaningful alerts than the historical rules-based kind of engines.” 

For $1.7 trillion Citi, existing AI tools that have helped keep a check on fiat money transactions will aid in tracking on-chain fraud as well, Southall said. The bank is continuously evaluating new tools to make its existing efforts more efficient, he said. 

However, Lenny Gusel, Head of Fraud Solutions at Feedzai thinks otherwise. Lisbon-based Feedzai is an anti-money laundering platform and TKTK. 

While an overwhelming majority of financial crime is still happening via fiat, digital currencies are one of the preferred avenues to steal money out of victims’ accounts, he told FinAi News. 

Different methods for different assets 

Financial institutions have had to add controls to transfer money to a cryptocurrency exchange, due to the high risk of fraud, and some have blocked the ability to make transfers to known cryptocurrency exchanges, Gusel said. 

“Banks can introduce degrees of ‘beneficial friction’ to increase the chances of catching successful fraud attempts, protecting would-be fraud victims from themselves in some instances,” he said.  

“Examples include how to use the data structure and format of a specific crypto asset or protocol, or how to understand the destination of a digital wallet address as opposed to a routing or account number.” 

For fiat money transactions, FIs must check bank account numbers, and other SWIFT details to avoid fraud, John Lunn, chief executive of payments provider Gr4vy, told FinAi News. “There’s obviously a layer of being anonymous that comes with crypto, so you need different rules.” 

Financial institutions increasingly offer tokenized assets stored in digital wallets, so the method of access to the wallets as well as interaction patterns with digital assets require different benchmarks for flagging and fighting fraud, Gusel said. 

Cross-institutional collaboration among FIs is also necessary to tackle the decentralized asset class, he said. 

AI tools to track patterns 

“AI emboldens bad actors to share information and tactics with each other more quickly and on a grander scale,” Gusel said. Therefore, “banks will also need trusted AI tools and omnichannel monitoring to maintain a holistic view of an individual’s or institution’s regular purchasing patterns and protect them in an emboldened threat landscape.” 

Feedzai is working with the European Central Bank (ECB), for example, to provide AML and fraud detection services for the digital euro, which is slated for launch this year or early 2027. 

The ECB will use Feedzai’s AI-driven RiskOps platform for financial crime prevention, Gusel said. 

The RiskOps platform provides customers with omnichannel, customer-centric monitoring, with rich digital channel devices and behavioral biometric signals developed to tackle fraud and AML for decentralized payment rails like crypto, Gusel said. 

Register here by Jan. 16 for early bird pricing for the inaugural FinAi Banking Summit, taking place March 2-3 in Denver. View the full event agenda here.