SPECIAL EDITION: BANKING, COMPETITION & CORPORATE LAW

 

Leveraging Artificial Intelligence to Combat Money Laundering and Related Crimes in the South African Banking Sector

 

 

H Chitimira; E Torerai; VLM Jana

 

North-West University, South Africa. Email: Howard.Chitimira@nwu.ac.za; elfas.torerai@gmail.com; lvjae@yahoo.com

 

 

---

ABSTRACT

Money laundering and related financial crimes, such as fraud and terrorism financing, pose a significant threat to the integrity and stability of South African financial markets. This article explores the application and use of artificial intelligence (AI) to detect and prevent money laundering in South African banking institutions. The implementation of big data technologies, data processing analytics and AI could enhance the detection and prevention of money laundering activities in South Africa's banking sector. AI should be carefully utilised to improve the detection of suspicious activities and the accuracy of financial intelligence, and to combat evolving money laundering techniques. The article also examines the benefits and challenges of implementing AI as an anti-money laundering (AML) measure in the South African banking sector. These include the need for quality data, integration with existing regulatory systems, regulatory compliance and ethical considerations. The article highlights the potential use of AI in transaction monitoring, customer due diligence, outcomes-based risk assessment and the improved detection of suspicious transactions. This could be done by utilising AI to enhance the effectiveness and efficiency of AML measures. The importance of effective coordination between banking institutions, regulatory authorities and law enforcement bodies is also highlighted as a key component of leveraging AI to combat money laundering and related financial crimes in South Africa's banking sector.

Keywords: Artificial intelligence; money laundering; banking institutions; customer due diligence.

---

 

 

1 Introductory remarks

Money laundering remains a persistent threat that is posing significant challenges to the integrity of global financial markets, especially in the wake of the technological challenges driven by global information and communication networks.¹ This necessitates the need to use advanced technology such as artificial intelligence (AI) to fortify anti-money laundering (AML) measures globally.² In 2015, the United Nations (UN) General Assembly established a worldwide framework to address global sustainable development, encompassing seventeen goals commonly known as the Sustainable Development Goals (SDGs).³ Each of these goals addresses specific social, economic, or environmental concerns, spanning from the eradication of poverty and the fighting of crime. SDG 16 is dedicated to "peace, justice, and strong institutions".⁴ This goal includes a sub-target 16.4, which specifically addresses the reduction of illicit financial flows and the combatting of organised crime.⁵ Target 16:4 provides that by 2030, there should be a significant reduction in illicit financial flows, strong mechanisms to recover and return stolen assets, and success in combatting all forms of organised crime globally.⁶

Money laundering and other related financial crimes represent a pervasive and escalating threat to the integrity and stability of financial markets, especially the banking sector.⁷ This is attributable, in part, to the intricate nature of modern financial transactions, coupled with the rapid evolution of money laundering techniques. Countries such as South Africa require robust and innovative measures in order to effectively combat money laundering and other related crimes.⁸ The Financial Intelligence Centre Act (FICA)⁹ provides some non-innovative measures to combat and curb money laundering in South Africa.¹⁰ The same status quo is duplicated in other AML regulations and statutes such as the Prevention of Organised Crime Act (POCA),¹¹ the General Laws (Anti-Money Laundering and Combating Terrorism Financing) Amendment Act,¹² the Protection of Constitutional Democracy against Terrorist and Related Activities Act¹³ and the Financial Intelligence Centre Regulations. AI is a handy tool for enhancing the efficiency and quality of financial services. However, it comes with risks to the integrity and financial stability of financial markets such as recording false positives and inadequately focussing on data quality assurance.¹⁴ For the purposes of this article, the advent of AI is seen in the broad context of advanced technological innovations that can be effectively used as techniques to enhance the effectiveness of AML measures in the South African banking sector.¹⁵ To this end, AI should be deployed to fortify AML measures in the South African banking sector.

However, the rapid evolution of the money laundering methods, trends, and techniques used by the perpetrators of money laundering offences has impeded the effective combatting of such offences.¹⁶ The current South African AML legal framework does not effectively combat the unique technology-driven risks and threats that are posed to market integrity and financial stability by the perpetrators of money laundering. This is owing to the fact that AI-driven financial transactions fall outside the scope of traditional AML regulations.¹⁷ This flaw could be exacerbated by the dynamic nature of AI systems, making it difficult for the current AML laws and regulations to keep pace with the emerging AI-induced money laundering techniques.¹⁸ Moreover, the lack of clear guidelines on the integration and use of AI by financial institutions to curb money laundering might create ambiguity and hinder the effective enforcement of AML laws, regulations and policies. The South African AML regulatory framework should be revamped to enact provisions that expressly address the intersection between AI and money laundering in order to provide guidelines on the responsible use of AI in the financial sector. These could include measures to ensure the transparency, accountability, and traceability of AI algorithms that are used in financial transactions, as well as the incorporation of AI-specific risk assessments into AML compliance programmes. Strengthening these aspects of the South African AML regulatory framework could enhance the combatting of all money laundering risks that are associated with AI in the banking sector.

Therefore, despite the potential benefits that AI brings to the financial sector, there are notable challenges and limitations that impede its reliability in the fight against money laundering and other financial crimes.¹⁹ The inherent complexity and adaptability challenges of money laundering schemes, which can evolve rapidly as a counter-response to detection methods employed by AI algorithms, affect the combatting of money laundering. Criminals may exploit vulnerabilities in AI systems by using sophisticated techniques to conduct illicit financial activities without detection.²⁰ Moreover, the reliability of AI models is contingent on the quality and quantity of the data which are received and processed by the relevant AI systems.²¹ If the collected data is biased or incomplete, the AI systems are likely to generate inaccurate or incomplete results which would ultimately lead to false positives or negatives in identifying suspicious transactions.²² Financial crimes often involve subtle patterns and intricate connections that may elude even the most advanced AI systems, particularly when dealing with the novel and/or sophisticated techniques employed by criminals. Additionally, the dynamic nature of financial markets poses a challenge for AI systems that may struggle to quickly adapt to emerging threats and regulatory changes through poor calibrations.²³ The poor calibrations and the complexity of some AI systems further complicate their integration into the AML framework, making it difficult for regulators and financial institutions to understand the reasoning behind AI-generated alerts. In addition, it should be noted that South Africa's AML regulatory framework does not expressly provide for the combatting of AI-induced money laundering and other financial crimes.

Accordingly, this article explores the use of AI-related AML measures to curb money laundering in the South African banking sector. The article also discusses shortcomings in the current South African AML regulatory framework. Thereafter, recommendations that could be utilised by policymakers to enhance the South African AML regulatory framework are provided.

 

2 The definition of key terms

The FICA defines money laundering as an act that involves the concealing or disguising of the nature, source, location, disposition or movement of the proceeds of unlawful activities in order to use financial resources derived from criminal activities as if they are from a legitimate and lawful source.²⁴This definition shows that money laundering could be described as a process designed to legalise illegal income or assets.²⁵ Money launderers always seek to disguise their illicit transactions to avoid detection. Money laundering is a multi-layered process which involves several transactions and participants to erase the true nature of the illicit funds which are laundered.²⁶ It also involves the commission of other crimes to acquire assets, resulting in the blending of these unlawfully obtained funds with legitimate sources, the obscuring of their origin, and the creation of an appearance of legitimacy at the conclusion of the process.²⁷ Although some commentators argue that the laundered funds should cross various jurisdictions in order to be considered truly "laundered", money laundering can still occur within a particular jurisdiction or country.²⁸ However, it is contended that it is not a fundamental prerequisite for money laundering since money may be laundered without leaving an economic area or crossing jurisdictions. Any money that is acquired through illegal means constitutes "dirty" money that requires a process of "cleaning" before its successful integration into the legitimate economy.²⁹ AI technologies could enable this cleaning transformation of dirty money to inadvertently empower those with such money to use it freely without fear of legal repercussions. Irrespective of the definition employed, the primary objective of money launderers is to mitigate or eliminate the risk of the funds being seized or forfeited after detection.³⁰ Ultimately, the overarching goal for money launderers is to enjoy the benefits of their crime without facing legal consequences.

Money laundering is a pressing global issue for regulators and law enforcement authorities. Those engaged in money laundering activities continually seek opportunities to legitimise their unlawfully gained assets.³¹Most forms of criminal activity inherently involve gaining and enjoying monetary benefits. Once criminals acquire cash and other assets from their illegal pursuits, they begin to engage in the process of money laundering to "cleanse" their illegal assets so that they appear legitimate.³² This is done to transfer illegally obtained money through legitimate people or legitimate accounts so that its original source cannot be traced. Money laundering allows wrongdoers to camouflage and transfer substantial amounts of illicitly gained funds. It empowers criminal enterprises to operate effectively.³³ The typical money laundering process involves three key stages, the first of which is placement, which occurs when illicit funds enter the financial system, often through depositing them into a bank account.³⁴ The second stage is the layering of the funds, which occurs when a series of transactions are completed with the intention of distancing the funds from their illegal origins.³⁵ Lastly, the integration stage occurs when the previously tainted funds are disguised as legitimate so that they are seamlessly blended into the financial markets.³⁶ Against this background, it then becomes imperative to discuss the role of AI in combatting money laundering in South Africa.

AI is an evolving concept that can be broadly defined as the use of powerful algorithms, machines and computer systems to simulate human intelligence, behaviour and capabilities through techniques such as machine learning and logic programming.³⁷ Using big data, the algorithms, machines and computers are able to perform human-like tasks. This is possible through, inter alia, identifying patterns and solving specific problems.³⁸ In this regard, AI stands a good chance of dismantling the intricate web of money laundering in South Africa's banking sector.

 

3 Overview background on money laundering crime in South Africa

Money laundering poses significant threats to the stability and integrity of financial markets globally. It is also a constant challenge to South Africa's banking sector. Pecuniary benefits serve as a key motivating factor for criminals when they undertake illegal activities such as human trafficking and money laundering.³⁹ For instance, human trafficking generates approximately US$150.2 billion per year globally for criminal organisations through activities such as money laundering, forced labour, sexual exploitation and organ harvesting.⁴⁰ A lot of dirty money is required to complete and further such criminal enterprises. In cases of human trafficking, money is needed to move the victims across various locations where they are exploited. As part of the criminal operations, money is necessary to bribe and pay the various complicit intermediaries who form part of the trafficking network that aids the criminal activities.⁴¹ Given the strong link between money and crime, most governments pursue initiatives to curtail the movement of money through money laundering among criminal organisations, in an attempt to reduce the criminals' incentive and ability to engage in illicit behaviour. South Africa functions as a prominent financial hub in Africa, boasting a robust banking and financial services sector along with a substantial cash-based market.⁴² This makes South Africa an attractive destination for criminals wanting to conduct illicit activities such as money laundering.⁴³ Although South Africa is historically known as a vibrant and diverse economy, it is also susceptible to money laundering and related crimes. In this regard, it is important to note that South Africa has a very low prosecution rate for serious financial crimes. This is partly because it has a weak primary AML legislative framework.⁴⁴ The low prosecution rate of money laundering crime implies that there is an ineffective application of the AML/CTF framework in South Africa.⁴⁵ Former Finance Minister, Trevor Manuel, estimated that South Africa lost between R21 billion and R85 billion through money laundering in the early 2000s.⁴⁶ It is possible that the level of money laundering has increased over the years.

Consequently, there is a need for a paradigm shift in the choice of AML strategies. AI technology should be recognised as an integral tool to use in effectively combatting money laundering in South Africa. While South Africa currently employs the traditional AML measures, there is a concern that they may fall short in combatting sophisticated money laundering threats that are technologically aided.⁴⁷ Hence, the rationale in advocating the use of AI technology-driven innovation in AML measures is underscored by the limitations of the existing rule-based systems in coping with the dynamic nature of money laundering.⁴⁸ This necessitates the adoption of a proactive and technology-driven approach to detect, prevent, and mitigate the risks associated with illicit financial activities such as money laundering. Moreover, the use of AI can produce measurable outcomes rather than following a limited rule-based approach which is increasingly becoming a tick-box exercise which does not effectively curb money laundering.

Against this backdrop, the integration of AI technologies into South Africa's AML regulatory framework presents a promising avenue to strengthen the resilience of banking institutions. The incorporation of AI technologies, particularly big data analytics, presents a paradigm shift in the fight against money laundering.⁴⁹ AI offers the capability to analyse vast datasets with speed and accuracy far beyond the scope and depth of traditional or manual AML methods.⁵⁰ AI systems can discern patterns, anomalies and trends which may be indicative of potential money laundering activities by leveraging machine learning algorithms. Thus, AI technologies will not only enhance the efficiency of AML processes, but they will also provide a proactive approach to identifying AML risks and emerging threats to the South African banking sector.⁵¹ In addition, the adaptability of AI technological systems allows them to evolve in line with the fast-changing nature of money laundering techniques, ensuring a sustainable and effective defence against such illicit financial activities.⁵²

 

4 The use of artificial intelligence to combat money laundering in South Africa

AI proves highly pertinent in this age of big data, given its capacity to process substantial amounts of data, including unstructured inputs like images and speeches.⁵³ AI measures such as machine learning algorithms are responsible for processing data inputs.⁵⁴ Machine learning utilises the computational process that aims to identify patterns in the dataset where the rules link inputs to outputs.⁵⁵ In addition, AI relies on predetermined rules that enable the computational process to apply these rules to input data in order to generate an output.⁵⁶ Nonetheless, AI and machine learning measures are not yet utilised in AML regulatory measures under the POCA and the FICA in South Africa.

Machine learning encompasses diverse types of computational processes, each tailored to address and solve specific problems such as detecting money laundering patterns in a particular jurisdiction.⁵⁷ The human supervised machine learning matches scenarios with known inputs and outputs.⁵⁸ In this regard, the human analyst manipulates identified money laundering detection datasets, separating and labelling data as either input or output. The AI algorithms detect patterns connecting inputs to outputs and formulate rules applicable to future instances of the problem.⁵⁹Conversely, unsupervised machine learning, where distinguishing between inputs and outputs is unknown, could give rise to mistakes and false conclusions.⁶⁰ The human analyst furnishes the computer with an unlabelled dataset, manipulating the algorithm by determining optimal data point groupings and formulating rules to elucidate their relationships.⁶¹ An intermediary technique, reinforced machine learning suits problems where specific actions yield superior results, like playing a game.⁶² In this context, the human analyst supplies the computer with a dataset, a goal and associated rewards or penalties for actions taken.⁶³ The algorithms should identify the optimal action to achieve the goal, sifting through potential data combinations and analysing rewards for different permutations to identify possible money laundering patterns.⁶⁴ The selection of the algorithm type should ideally align with the nature of the problem at hand.⁶⁵ However, in practice, this decision is often influenced by pragmatic considerations such as the human analyst's expertise, compatibility between programming languages⁶⁶ and/or the available processing power.⁶⁷ However, AI algorithms and related measures are still not used to detect and combat money laundering in South Africa under the POCA and the FICA.

Big data plays a pivotal role in the development of machine learning algorithms and the overall performance of algorithms. Without data, algorithms are likened to mathematical fictions.⁶⁸ Depending on the technical specifications, data can encompass structured data (numeric data) or extend to unstructured data like images or voice.⁶⁹ Datasets should be carefully utilised to detect suspicious activity that may be indicative of money laundering activities. These datasets are derived from historical databases such as shipping addresses or the types of internet protocol (IP) address connections which are used by a client when communicating with a financial institution.⁷⁰ Real-time data to analyse transaction patterns may be collected through physical sensors or online tracking or knowledge data such as the acceptance or rejection of previous financial product recommendations.⁷¹ Furthermore, data to create a clear client activity profile can be obtained internally or externally.⁷² Deciding which type of data to use or how much data to incorporate is often constrained by the necessity for systems compatibility across various elements of AI measures that are employed.⁷³ While the standardisation of AI technologies across institutions enhances the ability to utilise multiple and fragmented data sources for the regulator, it simultaneously diminishes the AI measures' flexibility and restricts contextual richness and competition amongst AI providers.⁷⁴Another critical consideration involves the quality of the harvested dataset, specifically how the data were gathered, their age and whether they are an accurate representation of the broader population's transactional patterns.⁷⁵Therefore, the challenge of the quality of the harvested data becomes pertinent in the case of externally controlled or acquired data, where firms may face challenges in accessing and evaluating the underlying assumptions and data sources driving an AML process for a financial institution.⁷⁶ Unfortunately, big data and other AI-related measures are not yet used by regulatory authorities to curb money laundering in the South African financial sector under the POCA and the FICA.

After the machine learning algorithm processes data for AML, the AML system generates an output whose nature and independence can vary.⁷⁷Machine learning autonomously acts based on computational results, like self-driving cars.⁷⁸ However, the system's output can be as straightforward as a score, lacking performative value until a human analyst acts upon it.⁷⁹This output can also be reintroduced into the AML dataset to further refine the algorithm's capabilities of detecting suspicious activity or unusual transactions such as money laundering.⁸⁰ This capability indicates that machine learning algorithms can learn and adapt over time to changes in their environment.⁸¹ Nevertheless, this adaptability can lead to complex and incomprehensible self-reinforcing feedback loops such as those generated by Facebook's AI negotiation bots, which create their own language.⁸²These loops can propagate biases and errors, as seen in AI-powered bots disseminating unverified information and automatic trading algorithms causing stock market flash crashes.⁸³ This issue becomes particularly significant in predictive analytics, where assessing output quality before implementation and scaling is challenging for human analysts.⁸⁴ AI and machine learning are capable of performing mechanical, analytical, intuitive or even empathetic tasks.⁸⁵ However, this is not yet being utilised in South Africa.

 

5 Leveraging AI to detect money laundering in the South African banking sector

Financial institutions should embrace AI technological measures to detect and prevent financial crimes such as insider trading, money laundering, terrorist financing and fraud in the South African banking sector. AI technological measures could enable banks to easily categorise customers and transactions as high or low risk so as to combat money laundering and related financial crimes.⁸⁶ AI algorithms automatically analyse transaction input and output to detect anomalies in the data provided to curb money laundering through predictive and predefined binary rules.⁸⁷ Computer scientists are making continuous efforts to enhance algorithm accuracy and reliability in order to reduce false alerts through meticulous record analysis.⁸⁸

AML measures follow a linear workflow in the banking sector, which is connected to a big data source.⁸⁹ In this context, human data analysts and programmers incorporate specific parameters to identify risky transactions, customers or communications which may suggest the occurrence of fraudulent behaviour.⁹⁰ A typical AML system comprises four layers, namely a data layer for collecting, managing, and storing data; a screening and monitoring layer for analysing clients and transactions for suspicious activities; an alert and event layer for raising alarms in case of suspected transactions; and an operational layer for further action.⁹¹ Financial institutions with effective AML regulatory frameworks rely not only on internal employee data but also on information collected from the fraud and sanctions watchlists of regulatory authorities.⁹² The data records undergo analysis using various AI and machine learning techniques such as natural language processing, enhancing computers' ability to comprehend and derive meaning from human languages and providing insights that establish connections between clients and transactions.⁹³

The screening and monitoring layer employs automated rule-based techniques using expert skills to identify potential money launderers.⁹⁴ Rule-based systems use pre-defined parameters and thresholds to detect suspicious trading activities. Striking a balance between being too strict (leading to multiple false alerts) and less strict (allowing illicit transactions), is crucial.⁹⁵ This layer retrieves user information from the data layer to screen customers and transactions. The alert and event layer signals the occurrence of a suspected transaction upon identifying suspicious activities.⁹⁶ Human AML analysts who are positioned in the first line of defence against money laundering in a financial institution then review the issues raised, taking manual actions such as allowing, rejecting, or blocking transactions.⁹⁷ However, the manual review process can overwhelm human AML analysts, creating backlogs in the client due diligence process, especially when the AI-powered system produces numerous false positives.⁹⁸

The effective integration of AI and data mining techniques will curb the challenge of numerous false alerts that could hinder the combatting of money laundering. Thus, the use of an outlier detection method enables the identification of fraud and/or money laundering activities on client profiles.⁹⁹This method empowers human analysts to establish various parameters for a client profile by scrutinising its transaction patterns. The analysis requires a vast dataset comprising hundreds of thousands of records spanning months to years, facilitating the classification of data into distinct clusters or groups. This approach has been acknowledged to be effective because it recognises similarities and differences in transaction patterns and classifies them into activity or pattern groups.¹⁰⁰ Once grouped as such, accounts or client profiles may be clearly differentiated between the levels of risks identified. Furthermore, given the evolving nature of money laundering techniques, AI can detect outliers and changes in behavioural patterns. Hence, AI and machine learning techniques excel in identifying suspicious transactions or irregular networks of money transfers not predefined in the outlier parameters.

5.1 Benefits of implementing Al in South African banking institutions

There are several benefits that AI can bring in combatting money laundering in the South African banking sector. Some of the benefits are discussed below.

5.1.1 AI enhances customer due diligence and know-your-customer measures

Effective data management is crucial for the streamlined delivery of financial services in the banking sector. AI is a key technology for enhancing data management, improving the speed, efficiency and accuracy of the services provided.¹⁰¹ Algorithms powered by AI can be developed to effectively detect and prevent money laundering and fraudulent activities by comprehending customer behaviours such as transaction frequency, amounts and network usage.¹⁰² Given the detrimental socio-economic effects of money laundering, banks should remain vigilant. AI technologies may play a pivotal role in proper customer identification to combat money laundering.¹⁰³

The primary defence against money laundering involves a thorough understanding of clients through the Know Your Client (KYC ) and Client Due Diligence (CDD) processes.¹⁰⁴ These processes entail that financial institutions should gather sufficient evidence to verify the identity and legal existence of individuals who seek to engage in a business dealing or entering into a business relationship with them.¹⁰⁵ The verification process includes scrutinising customers' official documents such as passports, identity cards, and proofs of names and addresses.¹⁰⁶ The CDD process enables banks to comprehend the nature of a client's business and sources of income to enable the effective monitoring of accounts for potential risks of money laundering. However, challenges arise in the digital era with electronic payment services and anonymous transactions. Therefore, there is a need for more robust and substantive outcomes-based approaches to AML, including harnessing AI technologies to detect illicit financial activities.¹⁰⁷

5.1.2 Integrating AI into the CDD processes of South African banks aids the detection of money laundering activities

Financial institutions such as banks collect customer data to assess loan risks, detect money laundering and combat fraudulent activities.¹⁰⁸ AI technology should be employed to enhance customer identification and verification and to promote the principles of CDD so as to curb money laundering.¹⁰⁹ KYC regulations are designed to aid financial institutions in monitoring and combatting money laundering globally.¹¹⁰ Autonomous AI systems help to process vast amounts of data, identify suspicious transactions, and detect patterns that could be indicative of money laundering transactions.¹¹¹

 

6 Challenges of using AI in the banking sector

Despite the positive acclaim around AI's efficacy in combatting money laundering, certain limitations impact its reliability. Malicious users can exploit AI systems to pose threats to both digital and physical security in the financial sector.¹¹² AI could pose a digital security threat through the poor use of machine and data analytics.¹¹³ Spear phishing attacks may be employed to gather critical details about individuals or to pilfer money or personal information, with attackers posing as trustworthy government or financial department entities.¹¹⁴ Therefore, AI can learn the specific habits of bank users to detect and curb money laundering. The accuracy of data sets in AML processes is crucial for AI to effectively detect money laundering and related illicit activities.¹¹⁵ Consequently, human interference can compromise the efficiency of AI-based AML systems through data poisoning attacks.¹¹⁶ The AI programs may learn from these manipulated inputs, enabling attackers to exploit the learned mistakes by introducing fake records during the calibration of AI systems.¹¹⁷ Given the automated nature of the process, it becomes challenging for humans to discern certain flaws, allowing money launderers to operate with impunity.¹¹⁸

Despite the obvious advantages of AI's big data analysis capabilities, implementing AI technologies in the South African banking sector to combat AML also presents multifaceted challenges. Firstly, there is a need for robust data infrastructure management. Many South African financial institutions struggle with fragmented data sources owing to their inability to cope with emerging technologies, making it challenging to build comprehensive AI models for detecting patterns of money laundering.¹¹⁹ A cohesive and integrated data framework is crucial for the effective functioning of AI-based AML systems.¹²⁰

Secondly, regulatory compliance poses a significant challenge to the combatting of money laundering in South Africa. The integration of AI in AML processes requires alignment with existing regulatory frameworks and ongoing collaboration between banks and regulatory bodies to ensure outcomes-based compliance while leveraging the benefits of AI.¹²¹Additionally, for an AML system to be effective, the AI that it employs should be capable of interpreting huge datasets accurately without being impeded by false positives inadvertently created by rigid rules-based regulations. The lack of interpretability can hinder substantive regulatory implementation by South African banking institutions and this may pose challenges in justifying AI-driven AML decisions to the regulatory authorities.¹²² To this end, the evolving nature of money laundering tactics requires the continuous adaptation of the AI-powered AML models and processes which are adopted by a financial institution. Criminals are constantly changing their strategies. Therefore, AI systems need to be agile enough to rapidly detect new money laundering trends, patterns and methods. This necessitates ongoing investment in research and development to ensure that AI-based AML systems remain effective against all money laundering activities in South African banks.¹²³

 

7 The utilisation of AI in the South African AML regulatory framework

7.1 Cybercrimes Act 19 of 2020

Money laundering is considered a cybercrime when criminals exploit digital channels, online platforms and digital financial systems to conceal the origins of their illicit funds.¹²⁴ In the context of cyber-enabled financial crimes, criminals may leverage various techniques such as online fraud, phishing, hacking or the use of virtual currencies to manipulate or infiltrate digital financial systems.¹²⁵ For instance, cybercriminals may engage in sophisticated schemes involving the illicit transfer of funds through digital channels, making detection and tracking very challenging for enforcement authorities.¹²⁶ The convergence of traditional money laundering practices and digital technologies underscores the need for the implementation of robust cybersecurity measures by financial institutions to prevent unauthorised access, data breaches and fraudulent activities.¹²⁷ As financial transactions increasingly occur in the digital space, the intersection of money laundering and cybercrime highlights the importance of making comprehensive efforts to combat both money laundering and cyber threats through the utilisation of AI in order to safeguard the integrity of the global financial system.¹²⁸ The Cybercrimes Act seeks to safeguard individuals, corporations, banks and financial institutions from cybercriminals, terrorists and individuals exploiting computers, the Internet and advanced technologies to commit cybercrimes in the country.¹²⁹ This Act prohibits any unauthorised access to data and the illicit acquisition of data by individuals. It seeks to address cyber threats and provide comprehensive protection against unlawful activities conducted through digital means.¹³⁰ Under the Cybercrimes Act, any person engaging in illegal interference with data or a computer program using either software or hardware tools commits a crime.¹³¹ This Act expressly prohibits activities that involve unauthorised tampering with, disruption of or interference with data or computer programmes, whether achieved through software manipulation or hardware tools.¹³² It seeks to establish legal boundaries and deter individuals from engaging in any form of illicit activities that compromise the integrity, security, or functionality of data and computer programs.¹³³

The Cybercrimes Act plays a crucial role in shaping the landscape for the use of AI in combatting money laundering in South Africa. It establishes a legal framework to address cyber threats that may intersect with financial crimes, such as money laundering, by explicitly outlawing illegal interference with data and computer programmes. AI is a powerful tool in analysing vast data sets and detecting patterns indicative of money laundering activities which can benefit from the legal provisions which are set out in the Cybercrimes Act. The Cybercrimes Act provides a foundation for the secure and lawful implementation of AI systems in financial institutions, allowing them to leverage advanced technologies to identify and prevent cyber-enabled financial crimes such as money laundering. This synergy between the Cybercrimes Act and the use of AI reinforces the regulatory environment, fostering a proactive approach to addressing emerging threats at the intersection of cybersecurity and financial crime prevention in South Africa. It is key to note that the Cybercrimes Act does not expressly provide for the use of AI to combat money laundering.

7.2 Electronic Communications and Transactions Act 25 of 2002

After years of legal ambiguity regarding the regulation of cybercrime, the enactment of the Electronic Communications and Transactions Act (ECTA) represents a crucial step towards combatting cybercrimes such as money laundering. The ECTA stands as the first legislative statute directly addressing cybercrime in South Africa.¹³⁴ While it refrains from explicitly defining "cybercrime", it provides a comprehensive definition of "access", covering actions where individuals, upon noticing data, are aware of their lacking authorisation yet persist in accessing it.¹³⁵ This statute prohibits unauthorised access to, interception of or interference with data.¹³⁶Importantly, these provisions not only establish a foundation for addressing cyber-related offences but also create an environment which is supportive of leveraging AI in AML measures in South Africa. AI has analytical capabilities and it can detect patterns in extensive datasets, making it instrumental in identifying instances of unauthorised data access, interception or interference, which aligns with the objectives of the ECTA. Furthermore, the Cybercrimes Act's prohibition of producing, selling or possessing devices or computer programmes which are designed to bypass data protection security measures is particularly relevant to ensuring secure AI implementation in combatting cyber-related offences. Additionally, the ECTA expressly addresses computer-related extortion, fraud and forgery, providing a legal foundation to combat cybercrimes.¹³⁷ The Cybercrime Act's provisions encompass individuals attempting to commit cybercrimes, reinforcing the legal framework's efficacy in addressing evolving threats in the digital landscape. However, the ECTA does not expressly provide for the use of AI to detect and combat money laundering.

7.3 ICASA and the Regulation of Al

The Independent Communications Authority of South Africa (ICASA) functions as an autonomous regulatory body which oversees the communications, broadcasting, and postal services sectors in the country.¹³⁸ It was established in 2000 through the ICASA Act. Its primary mandate is to regulate the telecommunications and broadcasting sectors in the public interest.¹³⁹ This role was assumed from the Independent Broadcasting Authority and the South African Telecommunications Regulatory Authority, a merger driven by the imperative to adapt to the rapid technological advancements occurring globally. ICASA is authorised to issue licences, monitor licensee compliance, and formulate regulations for consumer protection in terms of the Public Finance Management Act,¹⁴⁰ the ECTA, the Postal Services Act 24 of 1998,¹⁴¹ the Broadcasting Act 4 of 1999 and the ICASA Act. However, the ICASA and the ICASA Act do not specifically provide for the use of AI measures to detect, investigate, prevent and curb money laundering in South African banks and other financial institutions.

7.4 The POCA and the FICA

Although both the POCA and the FICA have some provisions that outlaw money laundering activities, they do not expressly provide for the use of AI measures to curb such activities. Thus, the POCA and the FICA do not require banks, financial institutions, enforcement authorities and other role-players to employ AI measures to combat money laundering activities in the South African banking sector.

 

8 Concluding remarks

The challenges to implementing AI technologies for AML processes in the South African banking sector underline the need for the adoption of a comprehensive and adaptable AI-powered outcomes-based approach to combatting money laundering. Despite its technological limitations, regulatory uncertainties and ethical considerations, there is a promising trajectory which is driven by potential positive shifts in both technology and AML regulatory framework in South Africa. Advancements in AI technologies, including the development of more sophisticated algorithms and enhanced machine learning models, hold the promise of improved capabilities in detecting intricate money laundering schemes and other illicit financial activities in South Africa. In this regard, the relevant regulatory bodies should be statutorily obliged to incorporate AI measures into their AML measures. Such measures should be clear and adequate. Moreover, regulators, financial institutions and technology experts should be statutorily obliged to collaborate and develop appropriate measures to curb money laundering in the South African banking sector. Banks and regulatory bodies should embrace and leverage AI technologies to combat money laundering and other financial crimes. The relevant AML statutes, such as the POCA and the FICA, should be carefully amended to oblige banks, financial institutions and other enforcement authorities to utilise AI measures to detect and curb financial crimes such as money laundering and fraud. Severe penalties should be imposed on all the perpetrators of money laundering and related financial crimes. Lastly, a robust and AI-AML integrated system that adapts to emerging threats and technological advancements should be adopted to safeguard the integrity of the South African financial sector against money laundering and related financial crimes.

 

Bibliography

Literature

Agarwal R and Dhar V "Big Data, Data Science, and Analytics: The Opportunity and Challenge for IS Research" 2014 Information Systems Research 443-448        [ Links ]

Alaimo C and Kallinikos J "Computing the Everyday: Social Media as Data Platforms" 2017 Information Society 175-191        [ Links ]

Alhajeri R and Alhashem A "Using Artificial Intelligence to Combat Money Laundering" 2023 intelligent information Management 284-305        [ Links ]

Allam Z and Dhunny ZA "On Big Data, Artificial Intelligence and Smart Cities" 2019 Cities 80-91        [ Links ]

Ball K et al The Private Security State? Surveillance, Consumer Data and the War on Terror (Copenhagen Business School Press Copenhagen 2015)        [ Links ]

Bara A and Le Roux P "International Financial Centres, Global Finance and Financial Development in the Southern Africa Development Community (SADC)" 2017 Journal of Economic and International Finance 68-79        [ Links ]

Basdeo V "The Legal Challenges of Criminal and Civil Asset Forfeiture in South Africa: A Comparative Analysis" 2013 AJICL 303-326        [ Links ]

Bhatore S, Mohan L and Reddy YR "Machine Learning Techniques for Credit Risk Evaluation: A Systematic Literature Review" 2020 Journal of Banking and Financial Technology 111-138        [ Links ]

Boles JR "Financial Sector Executives as Targets for Money Laundering Liability" 2015 Am Bus L J 365-433        [ Links ]

Byrne EF "Business Ethics Should Study Illicit Businesses: To Advance Respect for Human Rights" 2011 Journal of Business Ethics 497-509        [ Links ]

Calvard TS "Big Data, Organizational Learning, and Sensemaking: Theorizing Interpretive Challenges under Conditions of Dynamic Complexity" 2016 Management Learning 65-82        [ Links ]

Canhoto AI "Leveraging Machine Learning in the Global Fight against Money Laundering and Terrorism Financing: An Affordances Perspective" 2021 Journal of Business Research 441 -452        [ Links ]

Canhoto AI et al "The Role of Customer Management Capabilities in Public-Private Partnerships" 2017 Journal of Strategic Marketing 384-404        [ Links ]

Carlsen L and Bruggemann R "The 17 United Nations' Sustainable Development Goals: A Status by 2020" 2022 International Journal of Sustainable Development and World Ecology 219-229        [ Links ]

Chen TH "Do You Know Your Customer? Bank Risk Assessment Based on Machine Learning" 2020 Applied Soft Computing 1-7        [ Links ]

Chitimira H "An Exploration of the Current Regulatory Aspects of Money Laundering in South Africa" 2021 JMLC 789-805        [ Links ]

Constantiou ID and Kallinikos J "New Games, New Rules: Big Data and the Changing Context of Strategy" 2015 Journal of Information Technology 4457        [ Links ]

D'Amico G et al "Understanding Sensor Cities: Insights from Technology Giant Company Driven Smart Urbanism Practices" 2020 Sensors 1-24        [ Links ]

De Koker L "Money Laundering Trends in South Africa" 2003 JMLC 27-41        [ Links ]

De Koker L "Client Identification and Money Laundering Control: Perspectives on the Financial Intelligence Centre Act 38 of 2001" 2004 TSAR 715-746        [ Links ]

Enholm IM et al "Artificial Intelligence and Business Value: A Literature Review" 2022 Information Systems Frontiers 1709-1734        [ Links ]

Ennis J "Cleaning up the Beaches: The Caribbean Response to the FATF's Review to Identify Non-Cooperative Countries of Territories" 2002 Law and Business Review of the Americas 637-665        [ Links ]

Ferrara E et al "The Rise of Social Bots" 2016 Communications of the ACM 96-104        [ Links ]

Gates T and Jacob K "Payment's Fraud: Perception Versus Reality - A Conference Summary" 2009 Economic Perspectives 7-15        [ Links ]

Gaviyau W and Sibindi AB "Anti-Money Laundering and Customer Due Diligence: Empirical Evidence from South Africa" 2023 JMLC 224-238        [ Links ]

Gaviyau W and Sibindi AB "Global Anti-Money Laundering and Combating Terrorism Financing Regulatory Framework: A Critique" 2023 Journal of Risk and Financial Management 1-21        [ Links ]

Goecks LS et al "Anti-Money Laundering and Financial Fraud Detection: A Systematic Literature Review" 2022 Intelligent Systems in Accounting, Finance and Management 71-85        [ Links ]

Goetz AM and Jenkins R "Gender, Security, and Governance: The Case of Sustainable Development Goal 16" 2016 Gender and Development 127137        [ Links ]

Goodall NJ "Away from Trolley Problems and toward Risk Management" 2016 Applied Artificial Intelligence 810-821        [ Links ]

Goredema C "Confronting Money Laundering in South Africa: An Overview of Challenges and Milestones" 2007 Confronting the Proceeds of Crime in Southern Africa: An Introspection. ISS Monograph Series 73-92        [ Links ]

Hamman AJ The Impact of Anti-Money Laundering Legislation on the Legal Profession in South Africa (LLD/PhD-thesis University of the Western Cape 2015)        [ Links ]

Han J et al "Artificial Intelligence for Anti-Money Laundering: A Review and Extension" 2020 Digital Finance 211 -239        [ Links ]

Huang MH and Rust RT "Artificial Intelligence in Service" 2018 Journal of Service Research 155-172        [ Links ]

Huang H et al "A Survey of State-of-the-Art on Blockchains: Theories, Modelings, and Tools" 2021 ACM Computing Surveys 1-42        [ Links ]

Jana VLM A Comparative Statutory Analysis of the Regulation of Money Laundering in Zimbabwe (LLD-thesis North-West University 2024)        [ Links ]

Johari RJ et al "Money Laundering: Customer Due Diligence in the Era of Cryptocurrencies" Paper presented at the 1^st International Conference on Accounting, Management and Entrepreneurship (2020) 130-135        [ Links ]

Kersop M and Du Toit SF "Anti-Money Regulations and the Effective Use of Mobile Money in South Africa - Part 1" 2015 PELJ 1603-1636        [ Links ]

Kietzmann J, Paschen J and Treen E "Artificial Intelligence in Advertising: How Marketers Can Leverage Artificial Intelligence along the Consumer Journey" 2018 Journal of Advertising Research 263-267        [ Links ]

Khumalo K "The Problem with the Removal of the Motive Requirement from the Offence of Terrorism-A Short Commentary" 2023 African Security Review 1-5        [ Links ]

Kute DV et al "Deep Learning and Explainable Artificial Intelligence Techniques Applied for Detecting Money Laundering: A Critical Review" 2021 IEEE Access 82300-82317        [ Links ]

Labib NM, Rizka MA and Shokry AEM "Survey of Machine Learning Approaches of Anti-Money Laundering Techniques to Counter Terrorism Finance" in Internet of Things-Applications and Future: Proceedings of ITAF 2019 Singapore (Springer Singapore 2020) 73-87        [ Links ]

Marakalala MC and Mokwena RJ "Forensic Investigation: The Impact of Money-Laundering during Lockdown in South Africa" 2023 International Journal of Social Science Research and Review 523-532        [ Links ]

Menon S and Guan Siew T "Key Challenges in Tackling Economic and Cyber Crimes: Creating a Multilateral Platform for International Cooperation" 2012 JMLC 243-256        [ Links ]

Mirsky Y et al "The Threat of Offensive AI to Organizations" 2023 Computers and Security https://doi.org/10.1016/j.cose.2022.103006        [ Links ]

Mittelstadt BD et al "The Ethics of Algorithms: Mapping the Debate" 2016 Big Data & Society https://doi.org/10.1177/2053951716679679        [ Links ]

Ncube CB et al "Setting out the Challenges and Opportunities of Artificial Intelligence for Africa" in Ncube CB et al Artificial Intelligence and the Law in Africa (LexisNexis Durban 2023) 1 -20        [ Links ]

Nizovtsev YY et al "Mechanisms of Money Laundering Obtained from Cybercrime: The Legal Aspect" 2022 JMLC 297-305        [ Links ]

Njontini MN "Disruptive Technologies and the Future of Regulations: ICT Regulatory Structure(s) Determined" 2021 De Jure 174-193        [ Links ]

Paschen U, Pitt C and Kietzmann J "Artificial Intelligence: Building Blocks and an Innovation Typology" 2021 Business Horizons 147-155        [ Links ]

Pavlidis G "Deploying Artificial Intelligence for Anti-Money Laundering and Asset Recovery: The Dawn of a New Era" 2023 JMLC 156-166        [ Links ]

Roach K "A Comparison of South African and Canadian Anti-Terrorism Legislation" 2005 SACJ 127-150        [ Links ]

Russell SJ Artificial Intelligence: A Modern Approach (Pearson Education Inc Harlow 2010)        [ Links ]

Saha P, Bose I and Mahanti A "A Knowledge-based Scheme for Risk Assessment in Loan Processing by Banks" 2016 Decision Support Systems 78-88        [ Links ]

Schlenther B "The Taxing Business of Money Laundering: South Africa" 2013 JMLC 126-141        [ Links ]

Schlenther B "Is the South African Effort toward Reducing Money Laundering Optimal?" 2014 JMLC 17-33        [ Links ]

Silver D et al "Mastering the Game of Go without Human Knowledge" 2017 Nature 354-359        [ Links ]

Skiena SS The Algorithm Design Manual (Springer New York 1998)        [ Links ]

Sultzer S "Money Laundering: The Scope of the Problem and Attempts to Combat It" 1995 Tenn L Rev 143-238        [ Links ]

Truskauskas G and Taujanskaite K "Efficiency of Anti-Money Laundering: The Case of Northern European Countries" 2022 Business and Management 430-440        [ Links ]

Tuba MD "Prosecuting Money Laundering the FATF Way: An Analysis of Gaps and Challenges in South African Legislation from a Comparative Perspective" 2012 Acta Criminologica: African Journal of Criminology and Victimology 103-122        [ Links ]

Van Jaarsveld IL Aspects of Money Laundering in South African Law (LLD/PhD-thesis University of South Africa 2011)        [ Links ]

Whisker J and Lokanan ME "Anti-Money Laundering and Counter-Terrorist Financing Threats Posed by Mobile Money" 2019 JMLC 158-172        [ Links ]

Xu C et al "How Can a Blockchain-based Anti-Money Laundering System Improve Customer Due Diligence Process?" 2021 Journal of Forensic and Investigative Accounting 273-287        [ Links ]

Životić I, Ristić K and Mirković Z "The Impact of Security Management in the Detection of Money Laundering with Consequences on the Economy" 2022 Oditor 91-108        [ Links ]

Legislation

Broadcasting Act 4 of 1999

Cybercrimes Act 19 of 2020

Electronic Communications and Transaction Act 25 of 2002

Financial Intelligence Centre Act 38 of 2001

General Laws (Anti-Money Laundering and Combating Terrorism Financing) Amendment Act 22 of 2022

Independent Communications Authority of South Africa Act 13 of 2000

Postal Services Act 24 of 1998

Prevention of Organised Crime Act 121 of 1998

Protection of Constitutional Democracy against Terrorist and Related Activities Act 33 of 2004

Public Finance Management Act 1 of 1999

Internet sources

Bellomarini L, Laurenza E and Sallinger E 2020 Rule-based Anti-Money Laundering in Financial Intelligence Units: Experience and Vision http://hdl.handle.net/20.500.12708/58323 accessed 25 October 2023        [ Links ]

Brundage M et al 2018 The Malicious Use of Artificial Intelligence: Forecasting, Prevention, and Mitigation https://arxiv.org/pdf/1802.07228 accessed 4 July 2024        [ Links ]

Financial Action Task Force 2006 Trade-Based Money Laundering https://www.fatf-gafi.org/en/publications/Methodsandtrends/Trade-basedmoneylaundering.html accessed 20 October 2023        [ Links ]

Financial Action Task Force 2018 Financial Flows from Human Trafficking https://www.fatf-gafi.org/media/fatf/content/images/Human-Trafficking-2018.pdf accessed 5 June 2024        [ Links ]

Financial Action Task Force 2021 Opportunities and Challenges of New Technologies for AML/CFT https://www.fatf-gafi.org/en/publications/Digitaltransformation/Opportunities-challenges-new-technologies-for-aml-cft.html accessed 13 October 2023        [ Links ]

Financial Institutions Examination Council 2014 Bank Secrecy Act/Antimoney Laundering Examination Manual https://www.ffiec.gov/bsa_aml_infobase/documents/BSA_AML_Man_2014_v2.pdf accessed 20 October 2023        [ Links ]

Han J et al 2021 NextGen AML: Distributed Deep Learning-based Language Technologies to Augment Anti Money Laundering Investigation https://doras.dcu.ie/23358/ accessed 20 October 2023        [ Links ]

Hudson L 2017 Technology is Biased Too. How Do We Fix It? https://fivethirtyeight.com/features/technology-is-biased-too-how-do-we-fix-it/ accessed 20 October 2023        [ Links ]

Jacob K and Summers BJ 2008 Assessing the Landscape of Payments Fraud https://www.chicagofed.org/publications/chicago-fed-letter/2008/july-252 accessed 19 October 2023        [ Links ]

Jensen D 1997 Prospective Assessment of AI Technologies for Fraud Detection: A Case Study https://cdn.aaai.org/Workshops/1997/WS-97-07/WS97-07-007.pdf accessed 4 July 2024        [ Links ]

Khan I et al 2020 Algorithms: The New Means of Production https://www.digitalistmag.com/executive-research/algorithms-the-new-means-of-production accessed 30 October 2023        [ Links ]

Lewis M et al 2017 Deal or No Deal? Training AI Bots to Negotiate https://engineering.fb.com/2017/06/14/ml-applications/deal-or-no-deal-training-ai-bots-to-negotiate/ accessed 23 October 2023        [ Links ]

Mckinsey and Company 2022 The Fight against Money Laundering: Machine Learning is a Game Changer https://www.mckinsey.com/capabilities/risk-and-resilience/our-insights/the-fight-against-money-laundering-machine-learning-is-a-game-changer accessed 23 October 2023        [ Links ]

Mnih V et al 2013 Playing Atari with Deep Reinforcement Learning https://arxiv.org/pdf/1312.5602.pdf accessed 21 October 2023        [ Links ]

Moonstone Information Refinery 2022 Did Grey-Listing Threat Hasten SARB to Act against Markus Jooste? https://www.moonstone.co.za/did-grey-listing-threat-hasten-sarb-to-act-against-markus-jooste/ accessed 22 April 2023        [ Links ]

National Treasury 2023 Media Statement: Enactment of Key Anti-Money Laundering and Combating of Terror Financing Laws https://www.treasury.gov.za/comm_media/press/2023/2023010601%20MEDIA%20STATEMENT-ENACTMENT%20OF%20KEY%20ANTI-MONEY%20LAUNDERING%20AND%20COMBATING%20OF%20TERROR%20FINANCING%20LAWS%20.pdf accessed 5 June 2024        [ Links ]

O'Hear S 2016 Fraugster: A Startup that Uses A/ to Detect Payment Fraud, Raises $5m https://techcrunch.com/2017/01/16/fraugster/ accessed 19 October 2023        [ Links ]

Oseni A et al 2021 Security and Privacy for Artificial /ntelligence: Opportunities and Challenges https://arxiv.org/pdf/2102.04661 accessed 4 July 2024        [ Links ]

Raweh B, Cao E and Shihadeh F 2017 Review the Literature and Theories on Anti-Money Laundering https://scholar.ptuk.edu.ps/handle/123456789/639 accessed 20 October 2023        [ Links ]

Sanction Scanner date unknown What is Cyber-laundering? https://sanctionscanner.com/blog/cyber-laundering-and-cyberterrorism-494 accessed 5 June 2024        [ Links ]

Stervinou A 2015 Monitoring Payment Fraud: A Key Piece to the Puzzle https://www.kansascityfed.org/Payments%20Conferences/documents/7490/PSCP2015_StervinouPaper.pdf accessed 19 October 2023        [ Links ]

United Nations Department of Economic and Social Affairs date unknown Sustainable Development: The 17 Goals https://sdgs.un.org/goals accessed 5 June 2024        [ Links ]

United Nations Office on Drugs and Crime 2016 /ntroduction to Money-Laundering https://www.unodc.org/unodc/en/moneylaundering/introduction.html?ref=menuside accessed 14 October 2023        [ Links ]

List of Abbreviations

AI artificial intelligence

AJICL African Journal of International and Comparative Law

Am Bus L J American Business Law Journal

AML anti-money laundering

CDD Client Due Diligence

ECTA Electronic Communications and Transaction Act 25 of 2002

FATF Financial Action Task Force

FICA Financial Intelligence Centre Act 38 of 2001

ICASA Independent Communications Authority of South Africa

ICASA Act Independent Communications Authority of South Africa Act 13 of 2000

JMLC Journal of Money Laundering Control

KYC Know Your Client

PELJ Potchefstroom Electronic Law Journal

POCA Prevention of Organised Crime Act 121 of 1998

SACJ South African Journal of Criminal Justice

SDGs Sustainable Development Goals

Tenn L Rev Tennessee Law Review

TSAR Tydskrif vir die Suid-Afrikaanse Reg

UN United Nations

UNODC United Nations Office on Drugs and Crime

 

 

Date Submitted: 19 February 2024
Date Revised: 17 July 2024
Date Accepted: 17 July 2024
Date Published: 10 September 2024

 

 

Editor: Prof W Erlank
Journal Editor: Prof W Erlank
* Howard Chitimira. LLB (Cum Lauds), LLM (UFH), LLD (NMMU). Research Professor, Research Director and Professor of Securities and Financial Markets Law, Faculty of Law, North-West University, South Africa. Email: Howard.Chitimira@nwu.ac.za. ORCiD: https://orcid.org/0000-0003-1881-1242.
** Elfas Torerai. BSc (MSU), LLB (Unisa), LLM (NWU), LLD (NWU). Postdoctoral Research Fellow, Faculty of Law, North-West University, South Africa. Email: elfas.torerai@gmail.com. ORCiD: https://orcid.org/0000-0002-9680-5430.
*** Vimbai Lisa Michelle Jana. LLB (Wits), LLM (UWC), LLD (NWU), Faculty of Law, North-West University. ICA certified Money laundering Reporting Officer. Email: lyjae@yahoo.com. ORCiD: https://orcid.org/0000-0002-2509-8919. This article is influenced in part by Jana's LLD thesis titled: A Comparativs Statutory Analysis of the Regulation of Money Laundering in Zimbabwe (LLD-thesis North-West University 2024) 239-249. In this regard, I wish to acknowledge the expert inputs of Prof Chitimira and Dr Torerai.
1 Gaviyau and Sibindi 2023 Journal of Risk and Financial Management 13-14; Whisker and Lokanan 2019 JMLC 159-160.
2 Pavlidis 2023 JMLC 159; see related comments by Kute et al 2021 IEEE Access 82301.
3 UN Department of Economic and Social Affairs date unknown https://sdgs.un.org/goals; Carlsen and Bruggemann 2022 International Journal of Sustainable Development and World Ecology 229.
4 Goetz and Jenkins 2016 Gender and Development 127.
5 UN Department of Economic and Social Affairs date unknown https://sdgs.un.org/goals Target 16:4.
6 UN Department of Economic and Social Affairs date unknown https://sdgs.un.org/goals Target 16:4.
7 Van Jaarsveld Aspects of Money Laundering 7.
8 See related comments by Pavlidis 2023 JMLC 159-160; Menon and Guan Siew 2012 JMLC 245.
9 Financial Intelligence Centre Act 38 of 2001 as amended (FICA) ss 21-45; De Koker 2004 TSAR 718; Tuba 2012 Acta Criminologica 110.
10 See ss 20A-45 of FICA, which stipulate that financial institutions are obliged to do client identity verification in order to ascertain the identity of their customers.
11 See Prevention of Organised Crime Act 21 of 1998 (POCA) s 4; Kersop and Du Toit 2015 PELJ 1624; Tuba 2012 Acta Criminologica 109.
12 General Laws (Anti-Money Laundering and Combating Terrorism Financing) Amendment Act 22 of 2022 ss 18-20; National Treasury 2023 https://www.treasury.gov.za/comm_media/press/2023/2023010601%20MEDIA%20STATEMENT-ENACTMENT%20OF%20KEY%20ANTI-MONEY%20LAUNDERING%20AND%20COMBATING%20OF%20TERROR%20FINANCING%20LAWS%20.pdf 2; Moonstone Information Refinery 2022 https://www.moonstone.co.za/did-grey-listing-threat-hasten-sarb-to-act-against-markus-jooste/.
13 Protection of Constitutional Democracy against Terrorist and Related Activities Act 33 of 2004 s 2; see Khumalo 2023 African Security Review 1-2; Roach 2005 SACJ 130.
14 See Kute et al 2021 IEEE Access 82304.
15 Nizovtsev et al 2021 JMLC 299.
16 Enholm et al 2022 Information Systems Frontiers 1712.
17 Njontini 2021 De Jure 179.
18 See Njontini 2021 De Jure 179.
19 Alhajeri and Alhashem 2023 Intelligent Information Management 284.
20 FATF 2021 https://www.fatf-gafi.org/en/publications/Digitaltransformation/Opportunities-challenges-new-technologies-for-aml-cft.html.
21 Kute et al 2021 IEEE Access 82304-82305; also see Pavlidis 2023 JMLC 157-160.
22 FATF 2021 https://www.fatf-gafi.org/en/publications/Digitaltransformation/Opportunities-challenges-new-technologies-for-aml-cft.html.
23 FATF 2021 https://www.fatf-gafi.org/en/publications/Digitaltransformation/Opportunities -challenges-new-technologies-for-aml-cft.html.
24 Section 1 of the FICA.
25 Hamman Impact of Anti-Money Laundering Legislation 8.
26 The proceeds need not always be cash. They can assume any form, such as diamonds, gold, credit cards slip, stocks and bonds, cashier cheques, airplanes, rare coins, livestock, postal money orders, airline tickets, and wire transfers. Other synonyms, such as "tainted money", "illegal money", "hot money", or "black money" are also used to describe the proceeds of crime.
27 During this process the illicit proceeds of the criminal activities are intermingled with legitimate funds, which means that the "dirty money" is mixed with "clean money".
28 See related comments by Goredema 2007 ISS Monograph Series 78.
29 Goredema 2007 /SS Monograph Series 78.
30 See Goredema 2007 /SS Monograph Series 78.
31 UNODC 2016 https://www.unodc.org/unodc/en/moneylaundering/introduction.html?ref=menuside.
32 Boles 2019 Am Bus L J 365.
33 Boles 2019 Am Bus L J 365.
34 Gaviyau and Sibindi 2023 Journal of Risk and Financial Management 2-4; Ennis 2002 Law Business Review of the Americas 637.
35 Gaviyau and Sibindi 2023 Journal of Risk and Financial Management 2-4; Financial Institutions Examination Council 2014 https://www.ffiec.gov/bsa_aml_infobase/documents/BSA_AML_Man_2014_v2.pdf.
36 Gaviyau and Sibindi 2023 Journal of Risk and Financial Management 2-4; Sultzer 1995 Tenn L Rev 143.
37 Ncube et al "Setting out the Challenges" 1.
38 Ncube et al "Setting out the Challenges" 1 -2.
39 Byrne 2011 Journal of Business Ethics 498.
40 FATF 2018 https://www.fatf-gafi.org/media/fatf/content/images/Human-Trafficking-2018.pdf.
41 Ball et al Private Security State? 23-34.
42 See related comments by Bara and Le Roux 2017 Journal of Economic and International Finance 71.
43 Marakalala and Mokwena 2023 International Journal of Social Science Research and Review 529. British authorities have admitted that South Africa is being targeted as a money-laundering hotspot, and local authorities are fighting to make money laundering as difficult as possible.
44 See ss 4-8 of the POCA; De Koker 2003 JMLC 32; Basdeo 2013 AJICL 308-309.
45 See related comments by Chitimira 2021 JMLC 799; Schlenther 2013 JMLC 127.
46 Schlenther 2014 JMLC 21-22; Goredema 2007 ISS Monograph Series 78.
47 Gaviyau and Sibindi 2023 Journal of Risk and Financial Management 15-16; also see Goredema 2007 ISS Monograph Series 78.
48 Gaviyau and Sibindi 2023 Journal of Risk and Financial Management 15-16; Goredema 2007 ISS Monograph Series 78.
49 Mckinsey and Company 2022 https://www.mckinsey.com/capabilities/risk-and-resilience/our-insights/the-fight-against-money-laundering-machine-learning-is-a-game-changer.
50 Kute et al 2021 IEEE Access 82301.
51 Han et al 2020 Digital Finance 213.
52 Han et al 2020 Digital Finance 213.
53 Big data refers to extremely large and complex sets of information that traditional data processing methods struggle to manage efficiently. It involves the collection, storage, and analysis of vast amounts of data, often characterised by the three Vs: volume (large amounts of data), velocity (rapid data generation and processing), and variety (diverse types of data, such as text, images, and videos). Big data analytics involves using specialised techniques and technologies to extract valuable insights, patterns and trends from these massive datasets, enabling better decision-making and problem-solving; Kietzmann, Paschen and Treen 2018 Journal of Advertising Research 264.
54 Skiena Algorithm Design Manual 5.
55 Canhoto 2021 Journal of Business Research 442.
56 Canhoto 2021 Journal of Business Research 442-443.
57 See related comments by Kute et al 2021 IEEE Access 82304-82305.
58 Canhoto 2021 Journal of Business Research 442.
59 Canhoto 2021 Journal of Business Research 442.
60 Canhoto 2021 Journal of Business Research 442.
61 Kute et al 2021 IEEE Access 82304-82307.
62 Mnih et al 2013 https://arxiv.org/pdf/1312.5602.pdf.
63 See related comments by Kute et al 2021 IEEE Access 82304-82308.
64 Mnih et al 2013 https://arxiv.org/pdf/1312.5602.pdf.
65 See Skiena Algorithm Design Manual 5.
66 Calvard 2016 Management Learning 67.
67 See related comments by Agarwal and Dhar 2014 Information Systems Research 445-446.
68 Constantiou and Kallinikos 2015 Journal of Information Technology 46.
69 Paschen, Pitt and Kietzmann 2020 Business Horizons 150.
70 O'Hear 2016 https://techcrunch.com/2017/01/16/fraugster/.
71 Stervinou 2015 https://www.kansascityfed.org/Payments%20Conferences/documents/7490/PSCP2015_StervinouPaper.pdf.
72 Jacob and Summers 2008 https://www.chicagofed.org/publications/chicago-fed-letter/2008/july-252.
73 Gates and Jacob 2009 Economic Perspectives 7-15.
74 Alaimo and Kallinikos 2017 Information Society 175.
75 Hudson 2017 https://fivethirtyeight.com/features/technology-is-biased-too-how-do-we-fix-it/.
76 Khan et al 2020 https://www.digitalistmag.com/executive-research/algorithms-the-new-means-of-production.
77 Canhoto et al 2017 Journal of Strategic Marketing 386.
78 Goodall 2016 Applied Artificial Intelligence 810.
79 Silver et al 2017 Nature 354-359.
80 Silver et al 2017 Nature 354-359.
81 Russell Artificial Intelligence 5.
82 Lewis et al 2017 https://code.facebook.com/posts/1686672014972296/deal-or-no-deal-training-ai-bots-to-negotiate.
83 Ferrara et al 2016 Communications of the ACM 97.
84 Mittelstadt et al 2016 Big Data & Society.
85 Huang and Rust 2018 Journal of Service Research 159.
86 Allam and Dhunny 2019 Cities 89.
87 Alhajeri and Alhashem 2023 Intelligent Information Management 284.
88 Jensen 1997 https://cdn.aaai.org/Workshops/1997/WS-97-07/WS97-07-007.pdf 34.
89 Jensen 1997 https://cdn.aaai.org/Workshops/1997/WS-97-07/WS97-07-007.pdf 34.
90 Jensen 1997 https://cdn.aaai.org/Workshops/1997/WS-97-07/WS97-07-007.pdf 34.
91 Truskauskas and Taujanskaitè 2022 Business and Management 435.
92 Gaviyau and Sibindi 2023 JMLC 225-226.
93 Han et al 2020 Digital Finance 215.
94 Kute et al 2021 IEEE Access 82300.
95 Han et al 2021 https://doras.dcu.ie/23358/.
96 Labib, Rizka and Shokry "Survey of Machine Learning" 74.
97 Bellomarini, Laurenza and Sallinger 2020 http://hdl.handle.net/20.500.12708/58323134.
98 Goecks et al 2022 Intelligent Systems in Accounting, Finance and IVIanagement 7185.
99 Goecks et al 2022 Intelligent Systems in Accounting, Finance and Management 7185.
100 Truskauskas and Taujanskaitè 2022 Business and Management 430.
101 Životić, Ristić and Mirković 2022 Oditor 95.
102 See Životić, Ristić and Mirković 2022 Oditor 95.
103 Životić, Ristić and Mirković 2022 Oditor 95.
104 Životić, Ristić and Mirković 2022 Oditor 95.
105 Johari et al "Money Laundering" 130.
106 Xu et al 2021 Journal of Forensic and Investigative Accounting 274-275.
107 Raweh, Cao and Shihadeh 2017 https://scholar.ptuk.edu.ps/handle/123456789/639.
108 Chen 2020 Applied Soft Computing 1 -7.
109 Saha, Bose and Mahanti 2016 Decision Support Systems 78.
110 See related comments by Gaviyau and Sibindi 2023 JMLC 234.
111 Bhatore, Mohan and Reddy 2020 Journal of Banking and Financial Technology 111.
112 Brundage et al 2018 https://arxiv.org/pdf/1802.07228.
113 Oseni et al "Security and Privacy for Artificial Intelligence: Opportunities and Challenges".
114 Huang et al 2021 ACM Computing Surveys 8.
115 Mirsky et al 2023 Computers and Security.
116 Mirsky et al 2023 Computers and Security.
117 See Mirsky et al 2023 Computers and Security.
118 Mirsky et al 2023 Computers and Security.
119 Gaviyau and Sibindi 2023 Journal of Risk and Financial Management 15-16.
120 FATF 2006 https://www.fatf-gafi.org/en/publications/Methodsandtrends/Trade-basedmoneylaundering.html.
121 Huang and Rust 2018 Journal of Service Research 157.
122 Huang and Rust 2018 Journal of Service Research 157.
123 D'Amico et al 2020 Sensors 1 -24.
124 Sections 2-12 of the Cybercrimes Act 19 of 2020 (Cybercrimes Act); Sanction Scanner date unknown https://sanctionscanner.com/blog/cyber-laundering-and-cyberterrorism-494.
125 Sanction Scanner date unknown https://sanctionscanner.com/blog/cyber-laundering-and-cyberterrorism-494.
126 Sanction Scanner date unknown https://sanctionscanner.com/blog/cyber-laundering-and-cyberterrorism-494.
127 Sanction Scanner date unknown https://sanctionscanner.com/blog/cyber-laundering-and-cyberterrorism-494.
128 Sanction Scanner date unknown https://sanctionscanner.com/blog/cyber-laundering-and-cyberterrorism-494.
129 Sections 2 and 3 of the Cybercrimes Act.
130 Sections 4 and 5 of the Cybercrimes Act.
131 See ss 4 and 5 of the Cybercrimes Act.
132 Sections 4 and 5 of the Cybercrimes Act.
133 Sections 4 and 5 of the Cybercrimes Act.
134 Section 1 read with ss 10-89 of the Electronic Communications and Transaction Act 25 of 2002 (ECTA).
135 Section 1 read with ss 10-89 of the ECTA.
136 Section 1 read with ss 10-89 of the ECTA.
137 Section 1 read with ss 10-89 of the ECTA.
138 Independent Communications Authority of South Africa Act 13 of 2000, as amended (ICASA Act) ss 3 and 4.
139 See ss 3 and 4 of the ICASA Act.
140 Public Finance Management Act 1 of 1999.
141 Postal Services Act 24 of 1998.