Use of AI for AML Could Lower Effectiveness: Report

A new report from NICE Actimize and Regulation Asia sheds light on APAC’s use of technology to fight financial crime.

NICE Actimize and Regulation Asia have published a new report presenting findings from new research into the use of technology in financial crime risk management.

The research, based on survey and interview data collected from fraud and financial crime practitioners in Asia Pacific, finds that financial institutions (FIs) are showing an increasing willingness to leverage artificial intelligence (AI) and machine learning (ML) tools to boost automation and efficiency and achieve cost reductions.

According to the report, the most popular AI/ML use cases in financial crime risk management are to boost KYC processes, case management, investigations, transaction monitoring and screening systems.

The report suggests that FIs are spending much of their time testing and assessing new technologies prior to actual deployment in production environments, preferring to continue using more traditional or legacy systems until they can better understand and trust the new technologies, and secure regulatory support.

While the research finds that practitioners are adopting a cautious approach to the use of AI/ML, it also identifies that these capabilities are quickly maturing in the industry. Respondents to the research expect many FIs will be moving past POC projects this year and proceeding with larger scale implementations.

Efficiency vs effectiveness

The report warns that an overemphasis on the use of technology to gain efficiencies and cost reductions could potentially lead to lower effectiveness in financial crime risk management systems and processes, and higher overall risk.

Specifically, the report suggests that AI may be less effective at identifying novel or emerging risks that may not be present in historical training data, underscoring the need to still make use of human insight and oversight to ensure a resilient risk management framework.

“There is concern about whether the reliance on technology and automation to boost efficiency and save on cost may be reducing effectiveness, because we are relying less on people and the skills they have,” said Matthew Field, APAC Market Lead for Anti-Money Laundering at NICE Actimize.

“Technology is a massive opportunity, but … it could turn into a massive threat if we’re using it for an efficiency play rather than to improve effectiveness. Put simply, there is a risk that a focus on efficiency could make firms less effective overall, and ultimately expose them to higher risks.”

Fit for purpose

The report also features a Q&A with Carmen Chu, Executive Director at the Hong Kong Monetary Authority (HKMA). She discusses the regulator’s work to boost Hong Kong’s AML/CFT capabilities, including by encouraging the banking industry to work collaboratively and leverage data and technology.

On the use of AI, Ms Chu said banks should “carefully consider the nature of their AI applications and the level of risks involved”, and that ensure the auditability and explainability of such applications, use of good quality data, rigorous model validation, and effective oversight of third-party vendors.

The HKMA plans to issue guidance on the use of AI for AML/CFT this year to ensure such applications are ‘fit for purpose’, Ms Chu said, also highlighting the regulator’s own use of technology to assess sectoral money laundering risks.

The research also explored the prominence of certain thematic areas in suspicious activity reporting. Crypto-related activity was the area where the biggest increase in suspicious activity was seen compared to a year earlier, potentially due to “excessive caution” and “a degree of over-reporting”, the report says.

The full report is available for free download here.

Regulation Asia also recently hosted a webinar to discuss the initial findings of the research. The on-demand version is available here


To Top
Share via
Copy link
Powered by Social Snap