The crime of money laundering is one of the most important threats to the financial industry, therefore financial institutions equip themselves with technologically powerful and intelligent analytical tools to combat money laundering. Machine learning is one of the most effective ways in financial institutions’ anti-money laundering (AML) efforts. The algorithms operated by the machine learning technology are able to improve the identification methods to a great extent.

Regulatory Pressure about Anti-Money Laundering (AML)

It is fully under the responsibility of the financial institutions to detect unusual or suspicious transactions and report them, therefore they need to ensure that they employ a sufficient transaction monitoring and sanction screening system. According to most of the compliance teams machine learning is the perfect solution for anti-money laundering to minimize the regulatory pressure.

How can Machine Learning (ML) support Anti-Money Laundering (AML)?

1. Machine Learning (ML) in reducing the false positives by effective transaction monitoring and investigation of alerts

One of the common challenges of transaction monitoring and anti-money laundering (AML) processes is the generation of numerous suspicious activity alarms. It is estimated that only 1-2% of these alerts are real threats while the rest of 98% are false positives. Machine learning systems will support financial institutions in investigating these alerts and detecting and deactivating the false positives which eventually enables financial institutions to reserve more sources and time for the actual suspicious cases.

In addition to the capability of analyzing suspicious activities, machine learning (ML) systems can classify alerts into groups based on their risk levels: critical, high, medium and low so that the critical and high-risk alerts can be prioritized and acted upon rapidly.

2. Machine Learning (ML) in increasing efficiency in Know Your Customer (KYC) and Customer Due Diligence (CDD) procedures

The use of machine learning enhances anti-money laundering (AML) activities by improving the customer verification processes and detecting anomalies in customer behavior which directly make KYC and CDD procedures more efficient. Traditional monitoring and behavioral analysis models may remain behind the new technologies therefore they miss the new patterns as the money laundering criminals are generally one step ahead. Machine learning models on the other side are designed to detect any different or abnormal behavior in the customer by simply analyzing their actions during transactions.

3. Machine Learning (ML) in the analysis of unstructured and external data

The financial institutions need to analyze large amounts of data in order to ensure the successful Know Your Customer (KYC) and Customer Due Diligence (CDD) processes for anti-money laundering (AML) purposes. Implementing a risk-based approach is essential for the financial institutions to understand the personal, professional, social and political background of a customer. To reach out and evaluate this information, financial institutions should collect and analyze large amounts of data from media, public archives, social networks and open-source data sources. Although the traditional name searching methods may provide matches in external data, they will not be able to indicate the context or the possible relations with the politically exposed persons (PEPs). At that point, machine learning (ML), natural language processing and artificial intelligence (AI) technologies are of great help in analyzing the unstructured, big data and pointing out the connections and risks which is a really important feature for the anti-money laundering processes.