Ruling on use of AI in banking security puts Netherlands ‘ahead of the curve’

Out-Law News | 26 Oct 2022 | 8:56 am | 1 min. read

A decision by the Dutch Administrative High Court to allow a bank to use artificial intelligence (AI) in its transaction monitoring processes has been welcomed by one legal expert.

Wouter Seinen of Pinsent Masons said the ruling puts the Netherlands “back ahead of the curve” and offered more legal certainty that “deploying data-driven, AI powered solutions is in principle considered legitimate.” His comments came after the High Court ruled in favour of ‘neobank’ Bunq in its dispute with and the Dutch central bank (DNB).

The court said the central bank was wrong to ban Bunq from using AI in screening clients .The DNB had accused Bunq of failing to comply with both national and European anti-money laundering (AML) laws that require banks to determine risk profiles of new and existing clients and monitor their transactions accordingly.

The High Court found that the central bank had not proved that Bunq’s automated methodology for classifying clients in certain standard risk categories on the basis of “peer grouping” and its use of AI and data analysis did not comply with Dutch AML legislation. Banks usually rely on manual interviews and questionnaires to screen and categorise new clients into risk profiles. The court also rejected the central bank’s claim that Bunq lacked enough data on its business clients.

It held that the DNB should not make the use of manual processes mandatory for banks. The court said the Dutch law implementing the AML directives deliberately allows financial institutions to take a risk-based approach where possible to implement their own processes and controls within functional parameters. It also found, however, that Bunq had not done enough work to establish where clients’ money came from – including “politically exposed people”.

Seinen said: “This is a really important decision for Dutch banks because transaction monitoring and client intake and review processes are widely considered to be very costly and inefficient when they rely predominantly on manual efforts. This is certainly a territory where algorithms, big data and AI can reduce the cost of, and drive an improvement in the quality of AML and know-your-customer checks. I believe we are just in the beginning of this process and many new developments will arise”

We are processing your request. \n Thank you for your patience. An unknown error occurred, please input and try again.