Financial crime has existed throughout history, driven by the pervading pursuit of illicit wealth, but the advent of the internet has fueled an unprecedented wave of fraud, money laundering, and cyberattacks targeting financial systems. This article explores the enduring nature of financial crime and how the digital age has amplified both its scale and complexity.
The most pervasive and damaging forms of financial crime encompass a wide range of tactics, from identity theft and payment card fraud to complex investment scams and phishing schemes designed to extract sensitive financial data. Money laundering, perhaps the most widespread and well-known form of financial crime, is integral to the criminal ecosystem. By as it obscures the origins of illegal proceeds, facilitating their integration into the legitimate economy. Cybercrime poses a relentless threat, targeting financial institutions and individuals with data breaches, ransomware attacks, and other technology-based methods. Furthermore, the illicit use of finance underpins morally reprehensible activities such as terrorist financing, corruption that undermines institutions, sanctions evasion, and the vast criminal networks profiting from human trafficking and illegal trades.
These crimes inflict substantial financial losses on individuals, businesses, and entire economies while eroding trust in financial systems. Moreover, they generate profits that power further criminality and perpetuate cycles of harm. Combating financial crime requires a multifaceted approach emphasizing robust cybersecurity, enhanced due diligence for customer identification and transactions, and sophisticated data analytics to detect suspicious patterns. International collaboration between law enforcement agencies is crucial due to the transnational nature of these crimes. The stakes are high, making the ongoing fight against financial crime essential for protecting individuals, economies, and the integrity of our global financial systems.
Emerging Artificial Intelligence tools are increasingly being harnessed in the financial sector to combat financial crime. AI algorithms can analyze vast amounts of financial data rapidly and detect patterns indicative of fraudulent activities such as money laundering or fraudulent transactions. Over the last decade, with financial crime on the rise, many firms have started to employ AI tools such as Natural Language Processing (NLP), Risk Scoring and Behavioural Analytics, among others, to detect anomalies and flag potentially suspicious activities. Some of these firms include JPMorgan Chase, Barclay’s, HSBC, CitiGroup, Visa, Mastercard, and Bank of America. This article will further look at how HSBC, Visa and Mastercard have strategically employed AI tools in identifying and limiting financial crime.
AI is used in various ways to combat financial crime. For instance, machine learning models, a subset of AI can be trained on historical data to recognize complex patterns and trends associated with different types of financial crime. These machine learning models are typically adapted with new information and data which in turn helps improve their accuracy over time. Another example of an AI technique used to detect fraudulent activities is Natural Language Processing (NLP). These enable AI systems to analyze unstructured data and to extract information from them related to financial crime. Unstructured data comes in various forms – ranging from emails to news articles. This data can turn out to be very useful when detecting financial crime. NLP as an AI technique helps financial institutions stay informed about emerging threats and adapt their detection strategies accordingly. A third example of an AI technique used for fraud detection in financial services is network analysis. In this, AI algorithms work to analyze complex networks of financial transactions to identify connections between individuals involved in fraudulent activities. By further analyzing these relationships, AI can help identify fraudulent transactions and prevent financial crime.
HSBC is one of the most prominent financial institutions globally, with particular strength in Europe and Asia. Money laundering represents a significant threat to the integrity of the global financial system, with trillions of dollars in illicit funds circulated annually; HSBC, as a major international bank, has faced the challenge of effectively combating this criminal activity. Traditionally, the bank relied on a rules-based system to monitor transactions for signs of money laundering: while standard in the industry, this generated numerous false positives, resulting in wasted resources investigating ultimately innocent transactions. To address this, HSBC partnered with Google Cloud to develop an innovative AI-powered solution named Anti Money Laundering AI (AML AI). This system leverages HSBC's extensive knowledge of money laundering techniques to learn and identify suspicious activity autonomously, significantly increasing accuracy compared to manual optimization.
The results of AML AI's implementation have been impressive. The system demonstrates far greater precision in risk detection, identifying more suspicious activity while substantially reducing false positives. This translates to a more efficient use of investigative resources, doubling the amount of identified financial crime in commercial banking and quadrupling it in retail banking. Additionally, the AI's speed reduces the time needed to detect suspicious accounts, contributing to protecting the financial system from abuse. Importantly, AML AI can uncover criminal networks by recognizing patterns typical of money laundering, a task that rules-based systems struggled with. HSBC's success with AML AI was recognized with the Celent Model Risk Manager of the Year 2023 award, and the bank is now expanding AML AI globally, with its adoption by other financial institutions promising a significant collective impact in the fight against money laundering.
In addition to HSBC, two other examples of firms that have implemented robust fraud detection techniques using AI modeling techniques are Visa and Mastercard. More recently, Mastercard has launched a new generative artificial intelligence model to allow banks to better assess suspicious transactions on its network. This AI model claims to improve fraud detection rates by as much as 300% in some cases. This new advanced generative AI model – Decision Intelligence Pro – uses a cardholder’s visit as the prompt to determine whether the business involved in a transaction is a place the customer would likely go. This generates pathways through the firm's networks to find the answer – in the form of a score. A high score would follow the pattern of what’s the usual kind of behavior whereas a low score would indicate an anomaly. This process, with the help of generative AI algorithms, happens within just 50 milliseconds.
Over the last few decades, the financial sector has undergone a revolution thanks to digitalisation and the growing use of technology and AI. With this comes the increase in digital threats and fraudulent activities – which calls for efficient ways of dealing with, preventing and combating financial crime. In the future of cybersecurity, AI and machine learning models are more than likely to be further employed and harnessed to use algorithms to work towards financial crime prevention. For instance, we are likely to see further development in techniques like regulatory compliance automation which will help streamline data collection, analysis and reporting. Moreover, as the world starts to have a growing focus on ensuring the ethical use of AI in preventing financial crime and maintaining cyber security, we are likely to see increased investment in ethical AI frameworks to ensure transparency in the decision making process.