The ultimate price for non-compliance in the anti-money laundering (AML) industry could be death. There have been cases of banking irregularities that lead to the imposition of the death penalty for those involved. In the Western world, penalties are slightly more lenient, but appointed Money Laundering Reporting Officers (MLROs) are personally liable and could be imprisoned if there is proof that the preventions for money laundering were not sufficient. Criminals are becoming more sophisticated at laundering money by looking for loopholes in systems to avoid being caught. Terrorist financing, tax havens, migration trafficking and numerous other profit-generating crimes are linked to money laundering, making it imperative for financial institutions (FIs) to have suitable systems in place in order to prevent illicit money transfers.

In recent years the financial crime prevention industry has seen an increase in regulation driven by the behavior of the ‘bad guys’. Obscure crimes, lone-wolf terrorist attacks, human trafficking, increase in the use of digital currency and the global migration crisis are not only driving an increase in regulatory change, but also the pace of change. With the evolving nature of threats, recent regulation calls for faster response times to address financial crime; therefore, there needs to be a rapid acceleration in technological developments. So, is the technology such as Machine Learning a friend or a foe for compliance functions?

Terminology Disambiguation

Firstly, let’s make sure that we are aligned on terminology:

  • Artificial Intelligence – Artificial intelligence (AI) is a broader concept of machines being able to carry out tasks that we would consider smart. AI allows machines to learn from experience, adjust to new inputs and perform human-like tasks.1
  • Supervised Machine Learning – Machine learning (ML) is a subset of AI built on the idea that systems can learn from data, identify patterns and make decisions with minimal human interaction.2 Supervised ML utilizes algorithms, which already contain a known value for a target variable, where the humans provide answers to the questions during the training stage to the machine. This allows them to learn from humans in a ‘supervised’ fashion.3
  • Big Data – Big data is data that contains greater variety, arriving in increasing volumes and with ever-higher velocity. This is known as the three Vs… variety, volume and velocity.4
  • Data Analytics – Data analytics, more specifically as it relates to big data, examines large amounts of data to uncover hidden patterns, correlations and insights that can be extracted from the data itself.5

AI, ML, big data and data analytics in financial crime operations

Today, most people understand the data driven approach when it comes to their recent online shopping behaviour. We purchase goods online, and our preferred vendor will offer us goods that have previously been bought that might complement our purchase. The more data this algorithm has, the easier it is for it to make predictions and recommendations for users. Of course, this algorithm does not care about why the associations between recommendations exist, it only cares that they do. In compliance, industry analysts will care to know why.

FIs are focusing their efforts on introducing new technologies into their processes, and while there is still a lot of hype in the regulatory compliance market about them, the market research highlights four main areas that are being explored:

  1. AI and data analytics as an enabler of efficiencies for FIs;
  2. AI and data analytics as a cost cutting/labor saving innovation;
  3. Risk reduction through pattern detection; and
  4. AI and data analytics as a strategic value add for FIs.

How does all of this help you in your job?

The application of AI, big data and data analytics is not an easy task for FIs. As AI is pulling FIs in many different directions, they need to be clear on their investments and activities, and more importantly, they need to separate their marketing story from the implementation story.

As the volume of alerts, transactions and list entities grows with changes in regulatory environment, FIs and corporations are expected to absorb the rising operational costs for match remediation. Continually adding compliance staff only treats the symptom—too many false positives—and is not a practical, long-term strategy. Time spent remediating matches and distinguishing false positives from relevant risk drains resources and diverts attention from your core mission. Enabling efficiencies through the introduction of new technologies in your systems is one way to help eliminate the repetitive nature of level one analysts’ jobs.

Whilst there is a belief that the benefits of AI and ML are to cut human agents out of the process of on-boarding, it is important to understand that new technologies cannot fully replace humans. Technologies can help supplement and aid efficiency, but FIs need to think about repurposing their staff, rather than eliminating them. FIs should make human agents experts in exception handling in compliance, rather than agents that process volumes.