Artificial Intelligence and Machine Learning for Combatting Money laundering

Share

WMI partnered with NTU in a research project collaboration to evaluate and utilise various Artificial Intelligence and machine learning techniques to augment Anti-money Laundering (AML) prevention and detection capabilities. The objectives of this research project included to develop and apply an advanced AML which can better determine the relationship between clients’ probability of committing money laundering and their transaction patterns, and establish a technological infrastructure to enable systematic, timely and effective critical information and intelligence sharing among financial institutions in relation to this.

 

The research project identified existing gaps such as:

  • Lack of sophistication in relation to current datasets, assessments, technologies and systems for AML risk models
  • Lack of intelligence sharing measures and infrastructure for industry members to combat money laundering collectively.

 

Whereas fighting financial crimes such as money laundering is important and imperative, current processes are rife with inefficiency, misleading false positives, risks of false negatives, escalating investment costs with systems, technology and data challenges. Genuine clients are inconvenienced, business opportunities impeded, while bad actors go undetected.

 

This research project culminated in the development of an AI and machine learning based AML model, along with a software prototype that allows compliance and operations teams to import their data sets, train AI and machine learning models rapidly, and put these trained models to use to analyse new transactions, identify alerts based on probability of money laundering risk, prioritise investigations and facilitate STR reporting. The research output comprised an AML Framework and Model using Artificial Intelligence and Machine Learning encompassing an AML Industry Reference Model, AML model validation and a Security Architecture with Privacy Protection for Cross Bank Data Analysis.

 

We are glad to be able to work with NTU to integrate rigorous research and scientific expertise with real world industry application and contribution, through the support and partnerships with our Expert Panel comprising industry domain experts from UBS, DBS, UOB, OCBC, CAD and MAS. This has been an inspiring experience of collaborative spirit and public private partnership efforts to unite expertise to collectively combat money laundering as a community.

 

To download a full copy of the research papers, click here.

 

Although the research project has concluded, the journey to combat financial crime continues. Following the development of the AI and machine learning based model prototype, we welcome financial institutions who would be interested to know more or explore and trial run the software prototype for your organisation and environment. If you are interested in discussing this, please contact govaccred@wmi.edu.sg.