Federated Learning through Revolutionary Technology
The global anti-money laundering system is under enormous stress, with illicit actors still able to profit and launder trillions of dollars despite massive investment and efforts by financial institutions and authorities to prevent and track financial crime. There is a quest for greater effectiveness and efficiency through the use of new technologies like machine learning, while needing to secure and localize data and protect privacy. This paper presents a way of solving these policy objectives elegantly by leapfrogging the current system with a new design for the AML system. This model relies on using a federated learning architecture with DOZER technology from Consilient and Intel® Software Guard Extensions (Intel® SGX) technology to share insights into financial crime risks in a utility-like fashion. At scale, this new model can help securely and effectively discover systemically-relevant financial crime risk across institutions and borders, reduce the burden of false-positives and dependence on rules-based models, and protect privacy and security by moving the analytics and not the data.