Bunq’s win against the Dutch central bank (DNB) in the Netherlands’ court of appeal in October 2022 points to the challenges that regulators face in dealing with banks’ increasingly, non-traditional approaches to anti-money laundering (AML).
In October 2022, the appeal court ruled against DNB’s ban of Bunq after it had claimed that the latter failed to comply with AML laws by using machine learning (ML) techniques for screening its clients. While the court did find that Bunq’s practices came short of the required standards in terms of establishing sources of income and identifying how politically exposed individuals made their fortunes, it did not find DNB’s evidence sufficient to prove that Bunq had broken AML-related laws.
Go deeper with GlobalData
The court’s ruling has shed light on how ML and artificial intelligence (AI) have become critical components in banks’ compliance efforts. In a recent survey, more than half of financial institutions mentioned that they are already relying on AI solutions in their AML practices or were planning on adopting AI solutions within the next 18 months. This comes as no surprise given the benefits of using AI in terms of both cost reduction and enhanced performance. One company managed to decrease the number of false-positive transactions (i.e., transactions that are falsely flagged by legacy systems) by 90% or from 1,000 to 100 per day. This shows how ML techniques allow compliance officers to fully dedicate their time to high-risk transactions instead of wasting their efforts on false alerts.
Whereas DNB might not have had enough reasons to ban Bunq, calls for caution related to the use of AI in AML remain valid. To begin with, compliance officers could fail to spot regulatory breaches if they are not well acquainted with how machines are processing the data and on what basis transactions are being flagged. In addition, ML-based models can result in biases that could result in unfair outcomes if manual safeguards are not set in place. In Germany, one bank accidentally blocked the accounts of thousands of customers due to faulty, automated mechanisms put in place for AML purposes.
Despite DNB’s opposition to the use of AI in AML, lawmakers and policymakers in the Netherlands and elsewhere are updating their regulations to facilitate the use of AI in financial crime prevention. In the Netherlands, lawmakers recently proposed a bill that would allow for joint transaction-monitoring between different banks, which would increase the quality and amount of data available to individual banks and allow them to improve their ML techniques. Elsewhere, the UK and US governments have partnered to launch a prize challenge that encourages innovation in ML applications in AML, specifically those that harness multiple data sources without breaching privacy rules.
Ultimately, the adequacy of regulations will depend on how responsible banks are in their deployment of AI in their AML practices. Banks that ensure their relevant employees are well versed in how ML techniques function and selectively apply ML in areas where there is less risk of regulatory and privacy breaches (such as transaction monitoring) will both avoid conflict with regulators and develop a competitive advantage over more traditional banks. On the other hand, banks that adopt AI in AML without an appropriate awareness of the associated risks and proper training of their employees face the potential danger of legal battles with regulators and consumers and ending up with loss-making investments.
How well do you really know your competitors?
Access the most comprehensive Company Profiles on the market, powered by GlobalData. Save hours of research. Gain competitive edge.
Thank you!
Your download email will arrive shortly
Not ready to buy yet? Download a free sample
We are confident about the unique quality of our Company Profiles. However, we want you to make the most beneficial decision for your business, so we offer a free sample that you can download by submitting the below form
By GlobalData
