By 
A string of fines is shining a light on anti-money laundering failures at European banks.
First, the U.K.’s Financial Conduct Authority (FCA) slapped Gatehouse Bank with a £1.5 million ($1.83 million) penalty in October for not applying sufficient anti-money laundering (AML) checks on customers based in high-risk countries.
Then this month, the financial regulator hit Spanish bank Santander with a fine of over £107.7 million ($131.8 million) for AML failures relating to the way it monitored customer transactions.
And just last week, France’s financial regulator, the ACPR, disclosed that Crédit Agricole had been fined €1.5 million ($1.59 million) following an onsite inspection of the bank’s division Languedoc. The inspection uncovered shortcomings in the bank’s transaction-monitoring systems, customer due diligence processes and procedures for investigating suspicious payments.
The Crédit Agricole case is an example of financial regulators innovating to better enforce the rules.
In a press release relating to Crédit Agricole’s disciplinary proceedings, the ACPR stated that it was able to flag the bank’s failings thanks to an artificial intelligence (AI) model called LUCIA which was used during an onsite inspection for the first time in this case.
The tool allowed investigators to process nearly 70 gigabytes of data after the bank transferred the ACPR transaction records relating to 750 million payments carried out by customers.
With AI, the regulator was able to analyze payment details such as the currencies used and the location of the beneficiaries of transactions and other data, as well as cross-reference this information with customer due diligence files to identify suspicious activity.
Banks, AI Developers Team to Fight Financial Crime
While recent fines don’t look good for the banks involved, they don’t necessarily imply that their current AML technology isn’t up to scratch.
In fact, the ACPR’s use of AI to identify suspicious transactions is concordant with similar moves made by banks themselves, where AI technology is at the forefront of AML innovation.
In the case of Crédit Agricole, the bank is in the middle of a multiyear project to overhaul its AML tech stack and has partnered with Oracle to implement the tech giant’s AI models as part of its efforts to fight financial crime.
For its part, Santander has been deploying ThetaRay’s AI-powered analytics capacityas part of its AML architecture since 2020 to help detect money laundering schemes in correspondent banking transactions.
In the meantime, legal and judicial frameworks have had to adapt in response to financial institutions (FIs) putting AI at the center of their AML activities.
A prime example is Dutch neobank Bunq, which successfully sued Dutch National Bank (DNB) over its move to prohibit the digital bank from automating previously manual anti-fraud and AML processes using AI.
In the Bunq vs DNB case, the judge sided with the challenger bank’s position that AI-powered automation is not only allowed within EU law but can in fact be more effective than old-school, manual procedures.
Unraveling AI Complexities
In the world of AML tech, AI alone isn’t enough to prevent money laundering. Machine learning (ML) algorithms are applied to analyze large volumes of transactions and identify suspicious activity, as the ACPR did when it investigated Crédit Agricole.
However, making sense of the patterns AI observes is one of the main challenges AML professionals face today, Gudmundur Kristjansson, founder and CEO of AML software company Lucinity, told PYMNTS in an interview.
He added that prior to deploying Lucinity’s AI-powered solution, many of the company’s clients reported that their AML teams spent 80% of their time translating system outputs into meaningful information.
After all, AI can only deliver a statistical likelihood that a given transaction, account, or individual is involved in money laundering based on the information it has previously been fed.
The next stage of the process — confirming if something is truly suspicious — can only be automated so far, and human investigators are still needed to help banks comply with regulations and protect their customers.
Comments