The Financial Crimes Enforcement Network (FinCEN) in the United States just completed its request for information, gathering ideas from across the industry on how to better combat organized crime with innovations in artificial intelligence (AI) and machine learning as it seeks to modernize the Bank Secrecy Act and Anti-Money Laundering Regulations. 

Featurespace responded to FinCEN’s review of the Bank Secrecy Act (BSA) Regulations and Guidelines (Docket No. FINCEN-2021-0008 / 2021-27081), with recommendations across four key areas. 

1. Explainability  

Dispel confusion and create confidence in innovation programs through a clear, risk-based framework. Financial institutions (FIs) need to know what an acceptable level of explainability is as it relates to more advanced machine learning models. This would include defining explainability versus interpretability and guidelines on measuring performance and minimum effectiveness standards. 

2. Model Governance  

Update frameworks to be specific for financial crime management. Current guidelines were designed for a broader set of models, which makes it challenging for FIs to be confident that their innovation will be compliant. Regulation needs to ensure financial institutions can action the framework into practical solutions. 

3. Open Data  

Create APIs and collaboration groups to enable secure data sharing. The biggest challenge in AML/CTF and combatting organized crime is operating on minimal data. A more efficient, timely, and automated flow of information to FIs would provide richer data to machine learning models for additional accuracy in financial crime prevention. 

4. Structured Data 

Structure data to make it easily consumable for FIs to leverage in AML/CTF systems. Capitalize on new data standards like ISO 20022, mandate complete data fields in transactions, and align these to Suspicious Activity Reports (SARs) filings.  

With the right regulatory framework, machine learning models can be applied successfully to identify suspected money laundering and terrorist financing with greater accuracy and efficiency. Innovations in AI for AML hold the potential to cut off the illicit revenue streams that fund criminal networks and destroy their operations to prevent the trafficking of humans, drugs and illegal weapons.  

Read the complete recommendations from Featurespace here.