Policing Money Laundering With AI Offers Benefits and Challenges
Artificial intelligence (AI) can track billions of small transactions that would overwhelm human investigators. But this comes with a catch, as systems that are not well-balanced may detect too many or too few suspicious events.
Patrick Craig, financial crime technology lead at EY, a consultancy, highlighted that AI can enhance operational efficiency in areas such as customer due diligence, screening and transaction monitoring. However, implementing AI in anti-money laundering efforts is not without its challenges.
“The basis for developing AI-enabled [anti-money laundering] solutions must start with a clear statement of objectives to ensure that the design and implementation are aligned with the intended use and integrate effectively into business processes,” Craig said.
Money laundering involves the conversion of illegally obtained funds into legitimate assets, making it difficult for authorities to trace the origins of these illicit funds. With estimates suggesting that money laundering accounts for 2% to 5% of the global GDP, which amounts to $800 billion to $2 trillion annually, traditional methods of supervision and regulation fall short due to the sheer volume of transactions involved. And criminal organizations, terrorist groups, adversarial states and other nefarious actors depend on funding to sustain their activities that use both legal and illegal channels, thus crossing laundering lines several times.
Context is crucial when deploying AI in anti-money laundering. Simply implementing the technology without enhancing investigative and intelligence capabilities may yield limited results. This field requires effective human-machine interaction, as expertise and judgment are still essential for interpreting AI-generated alerts and making informed decisions.
“Establishing clear performance indicators and parameters, which link to a well-defined risk appetite statement, will be critical to tracking whether the outputs from the AI are meeting objectives at an acceptable level of risk,” Craig said.
The basis for developing AI-enabled [anti-money laundering] solutions must start with a clear statement of objectives to ensure that the design and implementation are aligned with the intended use and integrate effectively into business processes.
The question of whether these technologies will be used in businesses or governmental organizations is also relevant. Regulations surrounding anti-money laundering efforts vary worldwide, with private institutions often shouldering liability for accepting illegal funds. While there are no specific regulations for using AI in "know your customer" procedures, authorities still generally support firms experimenting with AI to strengthen compliance effectiveness, according to Moody’s Analytics, an investment intelligence company.
Despite the potential benefits of AI, there are technical challenges to address. Google reported that a significant number of system-generated alerts turn out to be false positives, leading to costly manual reviews and wasted investigation time. Google advocates for an adaptive approach, scoring transactions instead of relying solely on fixed rules, which could make it harder for wrongdoers to understand the parameters of enforcement.
AI holds promise in the fight against money laundering, but its successful implementation requires planning, clear objectives and ongoing adaptation. Beyond these challenges, potential benefits of increased efficiency and effectiveness in combating financial crimes make AI a tool in the future arsenal against global crime.