Standardisation of AML/CFT Regulatory Frameworks using Natural Language Processing

With the G-20 prioritising the development of a solution that will address the challenges of cross-border payment, a lot of focus and research has been exploring ways of using a decentralised ledger technology (DLT) solution to address this challenge. Some of the issues that have been identified with the DLT option are interoperability, scalability, security, and privacy. Central Bank Digital Currencies (CBDCs) implementing DLT have also been acknowledged by experts as having the potential to address some of these challenges. Given the focus on CBDCs as an alternative technology that could address the shortcomings of traditional cross-border payment systems, some of the design challenges that are impacting their widespread adoption have come to the fore. Considering that early CBDC projects were in reaction to emerging digital currencies and therefore were meant to address the domestic needs of each implementing jurisdiction, the lack of a common development framework for developing the technology contributed to the challenge of the interoperability of the various CBDCs across jurisdictions and has impacted it’s fullscale adoption as an alternative to existing cross-border payment systems. With specific focus on regulatory frameworks used by various jurisdictions and also an identified challenge of CBDCs, this paper explores Customer Due Diligence (CDD) regulation as a mechanism for combating anti-money laundering (AML) and counter-terrorist financing (CFT). As part of the efforts to adapt cross-border payment system into a world of distributed ledgers relying on smart contracts, it reviews the impact of divergent regulatory frameworks that limit the interoperability of CBDCs. It discusses some of the existing techniques in place for conducting CDD, such as Know Your Customer (KYC) and how they fit with the new technology. It concludes with a proposal on how the incongruence of legal frameworks and AML/CFT regulations, can be addressed and standardised using new technologies such as Large Language Models (LLMs).