Nov 16, 2021

Fintech Compliance: AI's New Role in Combatting Corruption

Fintech
Finance
corruption
Fraud
Antony Bellingall
5 min
Photo of a woman using computers to analyse data, depicting AI used for tackling corruption in financial services.
Discover how AI has started to be used as a tool for fighting corruption, what these technologies are, and why they are so important

Know your customer (KYC) and anti-money-laundering (AML) compliance have become a fact of life for all financial institutions. On-boarding new customers and maintaining existing ones requires a cottage industry of continuous monitoring and reporting using sanction lists (for example, HM Treasury list in the UK or OFAC in the US), developing a risk-based approach, monitoring transactions and reporting suspicions to regulators.

One of the more challenging AML requirements is to screen and identify politically exposed persons (PEPs). Such monitoring needs to occur because regulators deem PEPs to be high risk, and they will be subject to higher degrees of due diligence and checking than other customers. Screening such customers correctly, therefore, is a significant task.

This article will highlight the common challenges and potential solutions to screening PEPs.

What are politically exposed persons (PEP)? And why should we care about them?

There is no single definitive agreement on who is a PEP, but regulators around the world have tended to converge on a reasonably settled position. Essentially, a PEP has a high profile political role – they are in a position of some influence. Therefore, this would include politicians, senior civil servants, high court judges, senior members of political parties, and senior ranks in the armed forces. 

It gets slightly more nebulous as regulators also require the relatives and close associates (RCAs) of the PEPs to also be screened. Now how do we define these? Some regulators can be quite helpful here – the parents should be identified but not the grandparents, the siblings should be identified but not the cousins – or more simply put, we should look up to one degree of separation.

But then what about sons-in-law and daughters-in-law or ex-wives and husbands? On the close associate side, how close is close? A fellow director in a company may be considered close, but what if they just happened to work in the same company? Or did they act as a personal accountant ten years ago? Or they are members of the same club?

To help answer this, we need to understand why we should care about PEPs and RCAs in the first place. The simple answer is they present a higher money laundering risk due to the position they hold. This notion can be understood using two examples.

First, a politician or high court judge has greater access to the public money supply than a private citizen. The capacity and opportunity to perform criminal behaviour (even if in most cases they do not) is there.

Second, a high profile person like a politician can be subject to blackmail threats and other risks – and where there is blackmail, criminal money will usually lurk beneath. Only relatives or close associates who are close enough to the politician to also face these opportunities or threats should also, therefore, be correspondingly monitored.

As a result, PEPs need to be identified and subject to higher due diligence by financial institutions. This rule of thumb typically means deeper checking into where their money (source of wealth and source of funds) comes from as well as ensuring senior management sign off and approve their onboarding. 

What are the challenges with screening PEPs?

There is no single agreed global PEP list. This reality is not just because interpretations about them vary but also because they change so fast. Politicians come and go, generals and high court judges retire, marriages dissolve, and older relatives die. Keeping it all in one up-to-date list is a thankless task.

Moreover, PEPs themselves may be immediately difficult to identify if they don’t wish to disclose the fact that they are a PEP. Imagine if Elizabeth Alexandra Mary Windsor, or Alexander Johnson wanted to open an account with your company. Is either of these PEPs? Yes, the first is Queen Elizabeth II, and the second is more commonly known by his middle name Boris. So a PEP list not only needs to be constantly up-to-date, but it needs to know these alias names too. Though Her Majesty need not go through screening in the Commonwealth realm.

Moving on, while we may be able to identify a PEP, we then want to know what we should do with them if we find one. In most cases, the PEP or the RCA may simply just be high profile. Therefore, we just want to monitor them – but what if they were actually guilty of some corruption or fraud? We would like to prevent them from becoming our customers and report them to the regulators. So our search now takes on an added dimension; not only must we find PEPs, we must especially look for the bad apples among them too.

So how should we screen PEPs?

While the challenges are many, there are many factors that can help companies. First off, a risk-based approach is desirable. A national politician or a general in the army should be screened. A local councillor or sergeant in the army need not be – the categorisation of PEPs at the start of this article is designed to assist here. It’s all down to the degree of influence they hold.

Second, tools exist which can collect and synthesise data and provide real-time updates. This form of analysis gets around the issues of out-of-date information highlighted above. Developers will train such tools to review specific databases of politicians and use AI such as natural language processing (NLP) to hone the results automatically. 

And thirdly, such tools can be further tuned by looking at words and phrases that may indicate fraud or corruption – immediately bringing adverse media or negative news to the forefront.

At my company Idenfo, we have developed a tool to utilise context-based web parsing using NLP and deep neural networks to identify PEPs and RCAs from around the world and understand something about the adverse media context in which they may be engaged.

This development provides a self-learning optimised set of results providing real value to any financial institution. In simple terms, we deliver the results and the framework – this helps compliance teams use the data and make the appropriate decisions—another small weapon to use in the fight against financial crime and corruption.

***

Author: Antony Bellingall is Co-Founder and Director at Idenfo, a provider of compliance solutions and services to financial institutions around the world.

Share article