
IMF warns AI is reshaping financial stability risks and supervisory expectations
A July 23, 2026 IMF blog details how central banks and supervisors in Europe and Japan are already using machine learning on securities and derivatives data to detect anomalies, while institutions like the Federal Reserve, ECB, and Bank of Canada apply natural language processing to supervisory reports and consumer complaints. The IMF notes that most central banks plan to adopt generative AI for threat detection even as they recognize AI also strengthens attackers’ capabilities.
There is a strong case to invest in AI-based surveillance for trading, conduct, fraud, and cyber risk, potentially in partnership with regulators and industry utilities, to reduce loss rates and demonstrate leadership in safety and soundness.
highSupervisors adopting AI will have sharper tools to identify anomalies, creating higher expectations and lower tolerance for weak controls. Firms that do not modernize surveillance and risk detection may face more frequent findings and enforcement actions.
high