This is an authorised extract of the full publication where it refers to FinCrime Dynamics (formerly named EalaX). The full article can be found here: Aperio Intelligence Financial Crime Digest May 2021.

SPECIAL FEATURE

A closer look inside the UK Financial Conduct Authority’s Digital Sandbox

It is often said that governments and financial institutions are losing the battle against global fraud and money laundering. The United Nations Office on Drugs and Crime (UNODC) estimated that in 2009 around $1.6 trillion of criminal proceeds flowed through the global economy – roughly equal to the size of Canada’s gross domestic product. Meanwhile, trends of misinformation and disinformation online are requiring law enforcement and financial institutions to apply increasingly sophisticated technology to verify true information about individuals. The spectre of deep fakes – creating false depictions of real people - presents an additional problem, with AI enabling new forms of digital impersonation.

It is into this increasingly complex picture that innovative companies are fighting back, applying cutting-edge technology to a variety of problems. Over the past decade RegTech firms have flourished, creating practical applications of digital ledger technology, commonly known as blockchain, as well as biometric authentication and increasingly sophisticated financial crime analytic techniques. Several focus on “privacy-preserving” technology – an effort to solve the challenge of developing and using software in an industry which must treat data with the utmost confidentiality.

UK authorities have taken notice. HM Treasury’s recent Kalifa review of UK FinTech called for a “regulatory scalebox” to help start-ups grow. The UK government’s Economic Crime Plan 2019-2022 calls for HM Treasury and UK Finance, working with the

Financial Conduct Authority (FCA), Home Office, the Corporation of the City of London (CoCL) and other partners to establish a new Innovation Working Group to promote RegTech solutions to financial crime.

The FCA recently collaborated with CoCL on a ‘Digital Sandbox’ pilot, which provided support to FinTech firms tackling challenges caused by the Covid-19 pandemic. The pilot ran from October 2020 to February 2021, with 28 organisations participating. It provided a digital testing environment for the teams, with access to synthetic data, a coding development environment and a catalogue of FinTech and RegTech application programming interfaces (APIs). An observation deck enabled regulators to observe “in-fight testing” at a technical level. 

Several of the frms tested solutions designed to help combat fraud and scams. The Financial Crime Digest spoke to both breakout start-ups and established incumbents about their experience participating in the pilot.

EalaX – a synthetic data solution

EalaX founder and CEO Dr Edgar Lopez-Rojas is a Colombian data scientist and financial crime specialist who launched the company in 2019 to solve a specific problem, financial crime (Fraud and money laundering). Financial data held by financial institutions is highly confidential, but sharing it is crucial to building stronger controls and solving complex problems such as financial crime – or preventing it from even taking place. EalaX seeks to leapfrog this by generating synthetic data which mimics the behaviour of both licit and illicit transactions but avoids the confidentiality issues caused by the sharing of real-world banking information. 

Lopez-Rojas’ interest in financial crime was formed during his youth growing up in the city of Medellín in the 1980s. Notorious drug baron Pablo Escobar’s cartel was doing a roaring trade in cocaine between Colombia and the United States. “All this money ended up in my city. You might think that was wonderful, but the problems it caused to society in terms of corruption and violence were high. [Escobar] had so much power, he could manipulate society very easily. I saw the parents of my friends get killed across the street,” he said.

Lopez-Rojas developed an interest in computing, later moving to Sweden to pursue his PhD. During his studies he partnered with a company to apply machine learning techniques to money laundering controls. However, he lacked good data against which the controls could be tested. The experience led him to direct his research towards the simulation of synthetic data for financial services. 

Existing companies that provide software controls for financial institutions “do their best, but they are kind of blinded,” said Lopez-Rojas. “They don’t understand very well the modus operandi of the criminals. They are trying to do their best with the information they have, but the fight is unfair. The criminals have all the economic resources and go from institution to institution, finding the weakest link.”

Lopez-Rojas seeks to use synthetic data to avoid the confidentiality challenges that plague real data, thus building stronger connections between industry, academia and regulators. Regulators need better reports from industry, but they also need technologies which academia can develop. “Industry, in the financial sector and third-party vendors, has the potential to apply [the technology]. Sharing the knowledge is hard at the moment, but one of the ways to do this is through the use of synthetic data,” he said.

The FCA took an interest in his research, he worked as a consultant in 2019 and in 2020 signed EalaX up to provide one of the synthetic datasets for the digital Sandbox pilot. At the same time, the company secured two grants of £65,000 and £100,000 from Innovate UK, an arms-length public body which forms part of UK Research and Innovation. “We used the digital Sandbox as a platform for helping us to develop the technology,” he said. Prior to the Sandbox, EalaX didn’t have a product, but six months later the company is preparing to take its Synthetizor data solution to market. 

Lopez-Rojas said the Sandbox will play a “huge role” in the future of fintech. Of the response of Sandbox companies to his synthetic data solution, he said several were “saying good things… but at the same time they recognised limitations.” Lopez-Rojas acknowledged that the time-constraints of the Sandbox Pilot presented a challenge for developing all purpose generic data for participants working on a variety of fraud problems. “We had conversations with many of them, which are ongoing in terms of research,” he said. Following the conclusion of the Pilot, the FCA issued an evaluation report which found that the synthetic data was the most valuable Sandbox feature cited by participants while simultaneously the one with greatest potential for improvement.