Fraud Detection and Anti-Money Laundering

Overview

Leveraging artificial intelligence, the “Fraud Detection and Anti-Money Laundering (AML)” initiative develops tools capable of detecting fraud or money laundering patterns in real time. The models are designed to adapt quickly to emerging threats, while ensuring sufficient interpretability to meet regulatory requirements. This work is conducted in close collaboration with the compliance, cash management, and financial crime prevention teams.

Within this focus area, the theme “Fraud Detection for Cash Management/Transaction Banking” addresses a fundamental shift in transactional fraud. As digital payments grow, fraud can no longer be viewed as a series of isolated anomalies, but rather as coordinated and adaptive behaviors embedded within complex financial ecosystems. Detection approaches that remain largely rule-based or transaction-centric are increasingly out of step with this reality, particularly when fraud schemes are deliberately designed to blend in with legitimate activity. This research reframes fraud detection as a system-level intelligence problem. Rather than examining transactions in isolation, it models payment ecosystems as interconnected networks of customers, instruments, merchants, devices, and channels. This relational approach helps uncover organized, emerging, and concealed fraud patterns that cannot be detected through the analysis of individual events alone. The program specifically addresses the structural challenges faced by large financial institutions: the need to detect coordinated fraud in a context of significant class imbalance, to maintain performance as behaviors evolve, to integrate heterogeneous signals across entities, channels, and time, and to operate at scale with real-time, explainable decisions in regulated environments. By developing adaptive and interpretable network-based approaches that are compatible with operational constraints, this research aims to provide more robust, sustainable, and reliable fraud detection capabilities, while laying a solid scientific foundation for broader issues in financial crime, particularly the fight against money laundering.

Research Axis Leads

Joseph Gesnouin

Chief of Staff to Head of AI & Innovation for IT and Head of AI for Transaction Banking

Joseph Gesnouin is Chief of Staff to the Head of AI & Innovation for IT and Head of AI for Transaction Banking at BNP Paribas. He oversees the group’s AI industrialization strategy and its implementation, while managing the operational aspects of AI applications. Before joining BNP Paribas, he worked at the General Directorate of Public Finance (DGFiP) within the French Ministry of Economy and Finance. In this role, he advised the French Ministry of Finance on the use of language models (LLMs) to streamline budget preparation, combat fraud, and support other strategic activities. He holds a Ph.D. in engineering focused on AI applied to the automotive industry, earned at Mines ParisTech - PSL. His research has been recognized with various prestigious research awards, including the award for best young researcher in robotics from the National Center for Scientific Research (CNRS).

Aude Rousseau

Program Director of AML TM Tooling Strategy

Aude Rousseau, a French-qualified attorney and solicitor in England, worked for several years at law firms (Ashurst, Weil Gotshal & Freshfields). She joined BNP Paribas in London in 2016 as part of the internal audit teams, then assumed the role of Head of APAC AML TM, Standards & Procedures & Corporate Coverage in 2019. Initially based in Hong Kong and later in Singapore, she oversaw the teams responsible for the framework as well as the compliance investigation hubs for the 12 countries in the Asia-Pacific region—covering both CIB and Wealth Management activities. Aude Rousseau joined the Paris headquarters in January 2025 as Program Director of AML TM Tooling Strategy.

Yanlei Diao

Professor at École Polytechnique

Yanlei Diao is a professor of computer science at École Polytechnique and the University of Massachusetts Amherst. Her research focuses on big data analytics and scalable intelligent information systems, with a particular interest in explainable anomaly detection, stream processing, optimizing data analysis in the cloud, interactive data exploration, and managing uncertain data. She earned her Ph.D. in computer science from the University of California, Berkeley, in 2005.