The International Monetary Fund is urging banks to rethink a long-standing taboo in financial crime prevention—how much data they are willing to share with one another. In a new Technical Note, it argues that fragmented information is weakening the fight against AI-enabled fraud. combine data of their own, urging a shift in strategy toward sharing private information among financial institutions.
Released during the 2026 Spring Meetings of the IMF, the Technical Note focuses on how financial institutions can respond more proactively to digital fraud. The paper argues that efforts to combat such fraud have been hindered banks’ reluctance to share threat data, both domestically and internationally.
AI has become a boon for criminals, enabling them to aggregate vast amounts of data to fuel increasingly sophisticated attacks. In response, the IMF is urging banks to adopt a more collaborative approach—particularly by sharing more transactional and threat data cross institutions.
Importantly, the report cautions that new technology alone is not a silver bullet, warning against solutions deployed without clear use cases. Instead, it recommends robust data-sharing practices, especially around transaction records, to strengthen the collective ability of FIs to detect, prevent, and mitigate illicit finance activity.
More Data, More Value
AI and machine learning are highly effective at detecting transactional anomalies, but their performance depends on access to large, diverse datasets. When models operate in fragmented data environments, their insights are inherently limited. The Technical Note identifies these siloed data architectures as the primary obstacle in the fight against fraud.
By contrast, AI tools perform more effectively in integrated systems built on shared datasets, enabled by application programming interfaces (APIs) and common standards such as ISO 20022. The IMF highlights APIs, standardized data formats, and interoperability frameworks as essential to fostering meaningful data exchange across institutions.
Breaking Down the Silos
Banks have good reasons to resist data sharing, including competitive concerns and regulatory constraints. However, as fraud networks become more sophisticated and globally connected, greater transparency and collaboration could strengthen the financial system’s ability to detect and prevent illicit activity.
“Data sharing and collaboration is a long-standing issue within financial institutions, not even just between banks, but between lines of business within the same organization,” said Suzanne Sando, Lead Analyst of Fraud Management at Javelin Strategy & Research. “A financial institution may detect the signals needed to stop fraud on a customer’s credit card, but they may not be sharing the critical risk signals and emerging threat trends to stop fraud on that same customer’s debit account. These silos are preventing banks from accessing the critical data needed to keep up with fraudsters, especially as AI evolves and is adopted by fraudsters.”
