papersSEP 10 04:00 UTC
Paper models how a compromise at a shared AI vendor could cascade through the banking system
A newly posted arXiv paper examines the systemic risk created by banks' growing reliance on a handful of common AI suppliers for tasks such as fraud detection, credit decisions, and anti-money-laundering checks. It simulates how an attack or breach inside one of these vendors could propagate across institutions that appear independent, producing correlated failures throughout the financial system.