Whose cheque is it anyway? Why apostrophes matter for Anti-Money Laundering

Sabu Samarnath
Sabu Samarnath
3 min read

Google the word ‘check’ and you’ll receive two profound answers. One definition relates to the process of determining the condition of something, ensuring its quality and accuracy is in line with our expectations. The other: to halt the progress of an undesirable process. In the context of compliance and Anti-Money Laundering, both definitions are apt.

Growing up, the apostrophe presented me with a great deal of problems. Remember, in English grammar the apostrophe is used to indicate either ownership, or an abbreviation. For ownership, the apostrophe is placed after the name of the owner, and before the ‘s’. For an abbreviation, the apostrophe replaces the missing letters that have been removed. It should be so simple – right? This ‘when apostrophes gone wrong gallery’ proves otherwise.

This problem persisted to such a degree that my father started deducting pocket money for incorrect grammar. Through tears and trembling hands, my grammar steadily improved right up until the age of 18, when Microsoft Word and its heaven-sent ‘spell check’ function duly took over.

Fines are on the rise – and so are headcounts

My 50p fines imposed by Dad for mischecking work conjure up direct comparisons with the current state of play in today’s banking climate, with punishments from the FCA already reaching £350,630,287 for 2019 alone.

Whilst the sums of 50p and £350million are hardly comparable, I’m still pretty sure 50p had more effect on my bottom-line than theirs – and yet, unlike my switch to Microsoft Word, it seems banks are not turning to effective technology to relieve the pressure.

Instead, they’re recruiting extra people. Compliance and Fincrime personnel now equate to 3% of a bank’s headcount, a figure that’s doubled in 6 years according to Boston Consulting Group. Furthermore, banks spend $300 million annually staffing their AML operations, with an increasing reliance on off-shore and “compliance hubs”. HSBC’s chief compliance officer Colin Bell summarised it perfectly by stating “you have to build an industrial-scale operation just to digest all the regulatory changes.” An illustration of this systemic issue can be found in ING Group who, in the second quarter of this year, created 500 full-time positions to monitor suspicious transactions—a 20% rise, and a figure trumped by BNP Paribas with a 40% rise equating to 4,200 full time staff.
Going back to our definition of checks, it seems clear that in order for banks to consider transactions proper, and pull illegal activity to a halt, the checks they’ve decided to plump for put onus on a human workforce, an archaic methodology that will mean long training times, risk of human error, and increased costs. 

Scale knowledge, not resources

Rather than a scaling of human resources, banks should be considering a scaling of human knowledge via technology – something that can effectively replicate, en masse, the intricacies of human decision-making, allowing best-practice to always be the first line of defence. True, this type of paradigm-shift-thinking is going to require infrastructure upheaval, as current processes rely not only on more humans, but heavily on AI algorithms throwing up data for these humans to action. Tightening the parameters of black-box ML algorithms for spotting transactions might reduce the number of red-flag transactions, and thus the need for employees, but the combination of reduced margins and a lack of human oversight would also increase the risk of penalties from regulators.

Moreover, with the introduction of new regulatory frameworks centred around ethical AI – for example the new collaboration between the Alan Turing Institute and the FCA – it’s crucial that any new AML processes become explainable. Among regulators and consumers alike, trust and transparency in the use of AI has climbed to the top of the priority pile. Banks would be wise to follow suit and retain the human element in their AI-powered decision-making. As such, rules-based AI models are going to become crucially important. They not only scale human expertise effectively, but can also offer a level of accessible and affordable auditability that is simply absent in competing technologies. 

Transform your business into a Decision Intelligence powerhouse

Explore how Rainbird can seamlessly integrate human expertise into every decision-making process. Embrace the future of Decision Intelligence powered by explainable AI.