In today’s increasingly digital world, fraud and financial crime has become a significant concern. Global research shows that 2022 saw a 72% increase in fraudulent activity while almost a quarter of survey respondents expect a significant budget increase for anti-fraud technology through 2025. Given how generative artificial intelligence (AI) has been reshaping virtually all industry sectors, the challenge of fighting financial crime has become more complex and multifaceted.
Generative AI has been proven to be a groundbreaking technology capable of creating realistic data and media. But this means it has also opened significant new avenues for financial crimes. The increasing sophistication of fraud techniques, including deepfakes and synthetic identities, demands advanced detection and prevention strategies.
If anything, the world is on the cusp of what can be termed a “Dark Age of Fraud”. This will see the financial services sector scramble to employ AI solutions in a bid to counteract sophisticated fraud strategies. The space for positive use cases for generative AI is significant. Banks will look to invest in new technologies to counteract authorised push payment scams, forced by regulators for greater liability. Insurers will increasingly use this AI in their claims processes and fraud detection.
Generative AI also has the potential to transform fraud and financial crime compliance. By incorporating machine learning and network analytics into anti-fraud and anti-money laundering systems, the number of false negatives and positives can be reduced dramatically, increasing the efficiency of transaction monitoring.
Becoming more efficient
Therefore, to mitigate against the risk of the abuse of generative AI to perpetrate fraud, AI and machine learning must be used to enhance anti-financial crime programmes. Organisations can consider several strategies that will fundamentally change their approach to fraud detection.
At the most basic, they can leverage AI and machine learning to improve fraud detection accuracy and efficiency. Supervised machine learning algorithms can self-learn from target variables within the data, flag anything that does not fit the norm, and then apply this knowledge to new and unseen data. Unsupervised machine learning uncovers potentially suspicious types of risks organisations might not think to look for. It works without being given a target. Instead, it searches for anomalies in the data. Finally, entity resolution and network analytics can help to spot suspicious communities, and organised crime rings.
A second strategy encompasses fortifying and speeding authentication processes that validate customers in the digital world, leveraging multiple data sources related to intelligence, behavioural biometrics and checking the trustworthiness of information shared by customers. It can help to identify if we are dealing with real customers, with fraudsters, or with a bot. This can not only improve fraud detection but also reduce customer friction. Organisations can also consider using robotic process automation (RPA) to automate searches and queries of third-party data during enhanced due diligence processes.
The third aspect to consider is to coordinate and operationalise fraud, anti-money laundering, and cyber events. Given how financial services organisations are using big data analytics to consolidate data across siloed functions, it makes sense to combine these for a more holistic view of risk (called more and more often as FRAML). Much of the data and technology is similar, so the opportunity to reduce operational costs and enhance efficiencies cannot be ignored.
A fourth strategy is using AI to improve investigation efficiency with intelligent case management. An advanced analytics-driven alert and case management solution that presents a single view of data can automatically prioritise cases, recommend investigative steps, and fast-track straightforward cases. Additionally, this can intelligently find and pull data for a case from internal databases or third-party data providers while presenting data in easy-to-understand visualisations on one screen. That might be the future.
Responsible approach
When it comes to financial crime prevention, ethical considerations around AI must be a cornerstone. The intricacy of generative AI means that financial services organisations should not solely focus on the technological prowess but also on the ethical fabric that holds this technology.
Ensuring data privacy, securing informed consent where necessary, and preventing biases that could lead to unfair or discriminatory outcomes. Transparency in AI decision-making processes is crucial, allowing for auditability and explainability of AI-driven actions.
Next-generation anti-fraud and anti-money laundering technology has become imperative at a time when bad actors are using generative AI for fraudulent activities. As technology advances, the barrier to entry has dropped to the point where it is within reach of smaller institutions. Today, organisations do not have to have an army of data scientists on staff but can embrace packaged advanced fraud and financial crimes data science in a box to automate repetitive manual processes and more accurately detect suspicious activity.
Marcin Nadolny is the Head of EMEA Fraud, Fincrime & Data Science at SAS