AI in AML: Present tensed, but future perfect

384

September 9 2017

Today, Financial Institutions (FIs) face significant legal and reputational risks when it comes to complying with anti-money laundering (AML) requirements (including anti-terrorist financing and obligations to conform). Failure can lead to serious sanctions imposed by regulatory bodies (Recently, Societe Generale fined $5.83 MM for a number of shortcomings in its control for preventing money laundering).

AML systems are overwhelmingly rule-based. As regulations become more stringent, the rule-based systems grow more complex. Hundreds of rules drive know your customer (KYC) activity and Suspicious Activity Report (SAR) filings. As more rules get added, more cases get flagged for investigation and hence, false positive rates increase. The chase becomes ineffective. Criminal transactions (including terror funding) seem to outsmart the current system. However, unsupervised Machine Learning (UML) can address this by automatically detecting hidden patterns in large sets of data. AI in AML is still evolving. A more developed system in the future will save billions of dollars spent on inaccurate identification and chasing spurious transactions.