Artificial Intelligence for Trade-based Money Laundering Detection
Illicit trade is a problem of massive scale: Global Financial Integrity found a gap of US$8.7 trillion in reported trade between developing and advanced economies over ten years, yet the detection rate globally of illicit trade is near zero. We are facing a new age pandemic where $450 billion per year of illegally gained revenue is entering the genuine economy. If we do not act now, organised crime will soon direct all aspects of day to day life. Currently there are very few tools available to detect the complex, large scale Trade Based Money Laundering (TBML). Investigations are time and cost intensive. Law enforcement often relkes on tip-offs while customs has to do random checks. Globally we are lacking a science based understanding of how illicit transactions can be detected and what patterns they are following. The ATTENTION! project will analyse the largest trade database of imports and exports available globally and will use AI/ML models to derive patterns of illicit trade activity. Supervised and unsupervised ML models will identify patterns and train the models for increased accuracy.The trade activity of over forty countries over six years will be analysed comprehensively to ensure that all known smuggling methods such as co-mingling are included in the AI and ML models. The deliverable of the project will be an application to allow end users to perform checks on trade transactions. The application will return a risk rating. This tool will be useful to customs, law enforcement, banks, manufacturers, supply chain, hospitals to understand if the goods purchased, sold or financed could be compromised.