The effects of artificial intelligence on the AML landscape

As regulatory compliance initiatives move toward utilizing machine learning and artificial intelligence (AI), what effect might the increased efficacy of AML programs have on developing nations and small financial players without large compliance budgets? More specifically, will the eventual technology gap in the capability to detect and prevent money launderers create issues in shifting laundering risks or customer de-risking?

As AI matures, and large financial institutions (likely early adopters) incorporate it into compliance functions in order reduce costs, increase effectiveness, and manage volumes of customer and transaction data, a temporary gap in AML compliance capability will develop between the haves (global companies) and the have-nots (developing nations and local and regional entities). During the initial phase (gap stage) of AI development, the technology will be more expensive and cost-prohibitive to organizations with small compliance budgets. And if the purported effects of AI are true, the gap in AML compliance will quickly shift money laundering risks to smaller banks, credit unions, and payment services companies, as well as financial institutions in developing nations. The risk will shift because money launderers tend to operate along the path of least resistance. Illicit operations may also transition to less-detectable methods such as cash smuggling, precious metals exchange, digital currency transfer, and hawala.

What does this mean for the haves and have-nots?

The AI have-nots are likely to see the following effects during the gap stage:

• Increased exposure to regulatory focus, and reputational and operational risks
• Increased reliance and spending on updating legacy systems as “band-aid” solutions to short term risks
• Increased reliance on external compliance resources
• Additional transaction costs for customers
• Reduced access to the global financial system, especially in developing nations

The early AI adopters may also experience:

• Geographic de-risking of customers in developing nations due to less effective AML controls and the subsequent loss of revenue
• Hidden costs of managing new technology risks with regards to data privacy, customer experience, and immature regulatory expectations

How to bridge the gap?

There are some current non-AI AML initiatives being implemented globally that can help smaller organizations bridge the compliance gap. Biometric identification is a more mature (and currently cheaper) technology that has been leveraged to enable more effective KYC programs. There is also a variety of improved transaction monitoring systems available to smaller organizations, as well as some initiatives to centralize customer screening efforts at the government level. Even allowing AI systems to dig deeper into transactional data, then sharing identified risks with correspondent partners may optimize AI benefits across industry, though that would require an unprecedented level of cooperation between institutions. Leveraging these stopgaps and focusing on efficiency and process improvements in AML programs may be key in waiting out the prohibitive cost stage of AI implementation. While AI appears to be the next great solution to tackling money laundering, effectively managing the emerging and transitional risks of the technology learning curve will be key for the smaller players in the global financial network over the next decade.

Michael Carter is an expert consultant and thought leader in the area of Financial Crimes which includes Bank Secrecy Act / Anti-Money Laundering, OFAC programs, export compliance, and counter terrorism financing. His Twitter feed curates the latest in #AML, #ITAR, #FCPA, #EAR, and #WhiteCollar crime & #compliance news.