MAS to explore the full potential of AI to combat money laundering

During a recent interview with CNA (formerly Channel NewsAsia), managing director of the Monetary Authority of Singapore (MAS) Ravi Menon expressed a strong interest in exploring the full potential of artificial intelligence (AI) in the fight against money laundering. 

At present, MAS already utilizes machine learning and advanced data analytics to connect the dots across vast data sets to help detect fraud and other suspicious activities. However, according to Menon, a recent billion-dollar money laundering case highlighted the need for more to be done to strengthen “big-picture surveillance.”

Applying AI to COSMIC

One of the ways Menon suggested MAS could further utilize AI was in relation to the upcoming COSMIC digital platform, which is designed to facilitate the sharing of information on suspicious customers or transactions among financial institutions (FIs). Current confidentiality obligations prevent FIs from notifying each other about unusual customer activity. COSMIC, or Collaborative Sharing of Money Laundering/Terrorism Financing Information and Cases, aims to bridge this information gap when it becomes operational in the second half of 2024.

By integrating AI, and even potentially generative AI, into the platform, “additional insights” could be gained to create “a more comprehensive picture of the risks we face,” according to Menon.

The use of AI in the supervision of financial institutions

During a parliamentary session on October 3, 2023, Lawrence Wong, the Chairman of MAS, responded to a question about implementing AI in the supervision of FIs. The question related to whether AI has produced any positive outcomes for the regulator. Wong listed several benefits, including:

• MAS now has access to a wider range of tools and can better understand risk signals in the large amount of data it receives.
• Some manual tasks have been automated, resulting in increased efficiency and accuracy.
• Machine learning tools have made identifying outliers and suspicious networks easier, enabling closer scrutiny and more effective supervisory or enforcement action.
• Supervisors can now identify financial advisory representatives who present higher risks of illicit behavior.
• Natural language processing (NLP) technology has enabled supervisors to work more efficiently by analyzing texts and flagging issues for their attention.
• Social media and industry reports can be scanned by NLP-based tools, identifying news and developments that may require supervisory attention.
• Advanced data analytics, combined with AI and machine learning, can identify networks of suspicious activity in the financial system, allowing MAS to engage FIs and assess the robustness of their controls.
• These data analysis techniques and more powerful machine learning tools can help MAS sift out high-risk networks and transaction patterns more effectively.

Key takeaways for compliance staff

As forward-thinking regulators like MAS continue to explore how advanced technology can help disrupt flows of illicit funds, other regulators are likely to follow suit. To keep ahead of the fast-changing regulatory landscape, savvy compliance staff may benefit from engaging with topics such as those due to be discussed at the 2023 Singapore FinTech Festival. Taking place from November 15 to 17, the agenda highlights the following sessions relevant to the topics of AI and machine learning: 

• Ten artificial intelligence global insights.
• AI and existential change.
• Where are the real AI use cases?
• Ethical considerations for AI.
• Hyper-personalized experiences fuelled by customer AI.
• Revolutionizing finance with generative AI.
• Navigating GenAI: An executive roadmap for financial institutions.

Firms looking for guidance on how to use AI responsibly can refer to MAS’ Fairness, Ethics, Accountability, and Transparency (FEAT) principles, which aim to help compliance teams contextualize and operationalize the governance of AI and data analytics in their business models. For firms that aren’t ready for a full overhaul of their legacy technology, an AI overlay may be a more realistic alternative. Such solutions can integrate with existing tools to help prioritize risks and identify similar clusters of behavior that might otherwise have remained unseen.