Money laundering is a serious global problem. Despite this, there is little scientific literature on statistical and machine learning methods for anti-money laundering. Officials at the United Nations Office on Drugs and Crime estimate that money laundering accounts for 2.1-4% of the global economy. This figure indicates that the estimated amount of money laundered annually is almost 5% of global GDP, or $800 billion. Unfortunately, money laundering makes it difficult for honest businesses to compete because launderers are often able to offer their products and services at a cost below market value.
Considering the huge amount of money and social problems caused by this crime, it is justified to prioritize anti-money laundering (AML) efforts. Most of his AML regulations and recommendations now target cash transactions, which are most common in the early stages of the money laundering process. This type of banking transaction, which is usually conducted face-to-face, is subject to surveillance, unlike subsequent virtual transactions whose purpose is to prevent the tracking of funds through the various stages that make up the basic cycle of the money laundering process. Advantageous.
2022, according to McKinsey & Company. KPMG, 2018, The introduction of new technologies for anti-money laundering (AML) and combating the financing of terrorism (CFT) is hailed as a ‘game changer’ and ‘the coming revolution’. Emerging technologies in this area are often referred to as Regulatory Technology (RegTech) as they can be expected to improve his AML compliance and enhance delivery of regulatory requirements through AI.
OECD Principles for the Application of AI in AML
In 2019, the OECD developed guidelines to promote AI as an innovative and trusted technology that respects human rights and democratic values. The OECD Principles treat AI as a general-purpose technology rather than in a specific AML/CFT context.
First, it states that AI should promote inclusive growth, sustainable development and well-being (OECD Principle 1). This is a general and uncontroversial goal, but little concrete guidance. Second, it states that AI systems should be compatible with “the rule of law, human rights, democratic values, and diversity.”
A major challenge in this context is establishing effective mechanisms for informed surveillance. Informed regulation and supervision require the ability to develop expertise in new technologies, understand risks, and implement appropriate risk mitigation measures. AML supervisors must therefore be able to develop the expertise to effectively supervise the implementation of safeguards in AI-based systems by regulated entities.
The OECD Principles also call for greater “transparency and responsible disclosure about AI systems” to allow people to challenge results, which could affect decision-making by AI tools. This is important because it is not always understandable to potential humans. Financial institutions should not aim to eliminate the human component of AML. Rather, they should strive to support analysts with AI-based tools and free up resources for higher risk cases without abdicating responsibility.
EU AI law and its application to AML
AI forms the core of the EU’s Digital Single Market, and new harmonized rules are needed for the development and use of AI-based products and services in this market. The bill is the first comprehensive legislative initiative on AI by a major regulator to ensure transparency and accountability of processes and results, as well as respect for human oversight, cybersecurity, data protection and privacy. .
Core provisions of the proposed EU AI law define a risk-based approach to AI, categorizing AI-related risks as follows:
- Unacceptable risks such as social scoring prohibited by AI law.
- High risks, such as the use of AI in employee recruitment and medical devices. Permitted subject to prior third-party evaluation of suitability for AI requirements, taking into account laws in relevant areas.
- Low-risk AI systems with specific transparency obligations, such as chatbots impersonating humans, are permitted, subject to the obligation to notify humans that they are interacting with the AI system.and
- AI applications with minimal or no risk are allowed without restriction.
In the context of AML, the provisions of the proposed EU AI law also apply to the use of AI by regulated bodies and FIUs. The law and its risk categorization are compatible with FATF San Jose and OECD principles advocating for proactive and responsible AI innovation, fairness, transparency and accountability. The use of AI by regulated bodies and FIUs in the context of AML is considered high risk, as defined in Annex III subsections 6(e), (f) and (g) of the EU AI law. increase.
Conclusion
One hundred years ago, Nobel Peace Prize winner Christian Lange said: “Technology is a useful servant, but also a dangerous master. In the age of AI, this is a legitimate concern not only for nations, but also for the world’s largest corporations, including big tech companies and large financial institutions. The industry and the IT industry have already begun to revolutionize AML, where the use of AI can help detect and prevent money laundering by analyzing vast amounts of financial data and quickly and accurately identifying suspicious activity. Helpful.
