We teamed up with RegTech Associates to host a webinar around the topic of anti-money laundering (AML) and the role of artificial intelligence (AI) in advancing the efficiency and impact of the AML industry’s efforts to catch up to ever-evolving criminal typologies.
- Moderator: Dr Sian Lewin, Co-Founder and Head of Client Delivery, RegTech Associates
- Oliver Bullough, Journalist and Author of Moneyland
- Dr Janet Bastiman, Chief Data Scientist, Napier
- Mark Fungard, Senior Managing Director – Technical Sevices, K2 Integrity
The session began with Sian and Oliver discussing predicate crimes – crimes that generate illicit funds and precede money laundering – with a focus on corruption, Oliver’s area of expertise and the topic Moneyland. Kleptocracy and grand corruption have increasingly come into the media spotlight in recent years, and there’s growing awareness of just how widespread and harmful these crimes are.
Kleptocracy and money laundering
Oliver highlighted that, from a Western perspective, we may think of corruption as being something like the Mafia in New York or in Italy, as something that exists separately to the state, that the state fights against. In some countries, like the former Soviet Union, or most countries in the former Soviet Union, he asserts that for those who live there, it’s obvious that corruption is the state.
In these instances, where the state itself is corrupt, the term ‘kleptocracy’ is perhaps more appropriate to use – a word that describes a system whereby an entire country has been repurposed as a looting machine to enrich the people at the top and bring misery to the people at the bottom.
Where corruption and kleptocracy link to money laundering is that the proceeds, bribes and profits of these crimes are kept abroad, and must be moved and concealed to evade detection. More often than not, the UK is a stop along the way – or the destination - for this dirty money as it travels across borders.
“It’s still, I think, a mysterious phenomenon: the ability of very powerful people to essentially get away with anything by misusing the globalized financial system.” - Oliver Bullough, author of Moneyland
AML is more than just compliance
Sian highlighted that for those of us in working in AML, taking on money laundering, financial criminals, organised criminal networks, and corruption can be a frustrating undertaking, especially if there isn't the political will or systematic infrastructure behind you to make it easier.
She also pointed out that much of this frustration comes from the lack of efficacy in the fight against financial crime, and that many compliance professionals are all too aware of the very real harms that crimes like corruption, human trafficking or illegal wildlife trafficking cause every day. The crimes that AML professionals fight so hard to uncover are the sort that undermine democracy and cause human suffering – with the less fortunate always disproportionately affected.
We don’t talk enough about predicate crimes
To Janet, predicate crimes are incredibly important, and not spoken about enough. She explained that a lot of people can think of the money in isolation, or just within the context of a media story about the big banks getting fined for their involvement in money laundering scandals. Crucially, she pointed out that many people don't feel that it's something that affects them, and don’t stop to consider the underlying impact of financial crime.
AML is expensive and ineffective
Mark, in his numerous roles across law enforcement, regulation, and a financial institution, added that a tremendous amount of money is being spent, in institutions and in governments, to try to tackle the problem of financial crime in various ways. But despite the infamously high cost of AML compliance, unfortunately, the money spent has not yet led to a collective systemic approach or produced a way to actually solve the problem.
Mark also pointed to our understanding of typologies as a key area where AML is lacking. A number of controls, red flags, and typologies that persist today have their origins long in the past, for example, the anti-narcotics money laundering landscape of the 1980s. If you want to find 1980s cocaine money laundering in Miami, the controls are great, unfortunately, they hold little relevancy in catching the criminals of today.
Digitalisation is the future of AML
A lot of what we consider to be the standard typologies today, like cash-in-wire-out, come from that era also, because the technology when these early AML systems went into place was largely post Patriot Act in the early 2000s. This was a time when data storage and data processing was quite expensive, Mark recalled that it now costs mere pennies to store a gigabite of data today compared to around $10 in those early days.
Today, technology allows us the luxury of collecting and storing as much data as is needed, but as a community Mark felt that we haven’t yet caught up to technology and harnessed what it has to offer, especially the ability to gather large volumes of data to go after more sophisticated financial crime typologies.
The gap in typologies, as Mark explained it, comes from understanding how we get from a typology as laid out by law enforcement, to what that would look like represented in data and what the data features are that we can then use in AML systems to target it – something every institution struggles with to some degree.
For more about financial crime typologies, predicate offences, AML and AI, download our eBook: How AI can improve the detection of evolving typologies driving financial crime.
AI can help close the gaps in AML
Historically, the wrong data has been captured – and sparsely. Behemoth systems in big organisations, as Janet called them, are often very siloed and don’t allow for data to be connected up without significant effort. Criminals are very entrepreneurial in many ways, and recognise that there’s large gaps in these institutions that can be used to their advantage to evade exposure.
Janet reflected that one of the first things she does when talking to financial institutions is look at what they already have, because quite often the data has to be reorganised and stripped back to the basics before going in and implementing sophisticated new technologies like AI and machine learning – something which isn’t as big a task as it sounds but is crucial.
Ultimately, technology can overcome the historical issues data has presented in detecting suspicious activity or suspicious individuals by providing a lens through which to efficiently evaluate the data in one view. Janet also impressed that fundamentally, to get the best out of the latest technologies, you have to start at the beginning by gathering the correct information, because you can't predict or analyse something that you just don't have access to.
AI and information sharing
Increasing public/private collaboration represents some of the ‘green shoots’ of the beginnings of help, according to Mark. In the past few years, we’ve seen more public/private partnerships, primarily between financial institutions and law enforcement, where it’s been recognised that disparate, disorganised data presents a significant barrier to AML efforts.
Thankfully, emerging technologies such as federated machine learning and homomorphic encryption provide a new hope for the future of a secure, collaborative approach to data, and negate a lot of the heavy lifting associated with data sharing between separate entities.
AI and its potential to transform the current success rate of AML efforts globally is something which the panel agreed presents a massive opportunity for any government to take advantage of.