The use of AI in fighting corruption and financial crimes

This research focuses on identifying economic and financial crimes using statistical methods, machine learning, and artificial intelligence, particularly with regard to corruption. While economic crimes threaten markets and institutions more broadly, corruption is especially harmful as it diverts public resources, reduces the quality of essential services, and undermines citizens’ trust in the government.

The results aim to support policymakers and oversight institutions in making public spending more transparent and efficient by developing AI-based tools to detect suspicious patterns and risk factors. Tackling corruption helps safeguard taxpayer money and contributes to fairer competition, stronger institutions, and better outcomes for society as a whole.

Research method

The research involves combining different sources of data, such as information on companies, ownership, or public procurement contracts, to reveal hidden patterns that may indicate corruption or other financial crimes. Such crimes are often difficult to detect because they rely on complex corporate structures and networks of interconnected individuals.

To address this challenge, we will employ network-based machine learning and AI methods, which focus on the relationships between entities rather than their individual characteristics. For instance, modelling companies and directors as nodes in a network and their connections as links enables AI to recognise unusual structures or suspicious patterns that could indicate illicit behaviours.

This approach is powerful because, while corruption rarely leaves direct evidence, it does leave traces in data relationships. By systematically analysing these connections across large datasets, this research provides regulators and policymakers with new tools to identify risks earlier, monitor markets more effectively, and ultimately strengthen the integrity of financial and public systems.

Researcher

Andrea Longobucco, PhD candidate

Academic supervisor(s)

• Dr. Joras Ferwerda, Utrecht University School of Economics
• Prof.dr.ir. Marcel Boumans, Utrecht University School of Economics
• Prof.dr. Brigitte Unger, Utrecht University School of Economics