Today’s criminals have more sophisticated ways to launder money than ever before. This is creating unprecedented challenges for the AML community. To minimize the increasing risks, financial institutions also need more advanced tools and skills to help them track, discover, and stop financial crimes from occurring.
One new field of technology being explored in the fight against financial crimes is artificial intelligence. Through ongoing continuous training, machine-learning-based AI continually gathers new knowledge, becomes smarter, and adjusts processes according to new data inputs.
We’ve heard from many FIs who are wondering how AI could help support their anti-money laundering activities. While the capabilities of AI are growing rapidly, our industry is in the early stages of exploring potential opportunities and applying them to AML practices. Let’s review where we are and where we’re headed.
AML Opportunities with AI
Despite the vast resources already deployed by financial institutions to combat money laundering, financial crime is rising. Money laundering activity around the world is estimated to be between 2% to 5% of global GDP, according to the United Nations Office on Drugs and Crime. With billions at stake, criminals will never stop. In fact, money laundering is one of the key engines sustaining global criminal enterprises.
AI shows promise as a sophisticated tool to help financial institutions overcome common data-related issues that continually disrupt AML programs. AI may help solve poor-quality data, high rates of false positives in risk scoring, inconsistent customer identification processes, delays in suspicious activities reporting, and fragmented systems.
What’s more, it can bring advanced functionality to enable sophisticated AML capabilities, such as fraud identification, transaction monitoring, sanctions screening, know your customer (KYC) checks. It may also help with the integration of AML silos to provide a contextual basis for detecting suspicious activity.
Using the AI technology currently in development, the following scenarios may be possible to help with future AML efforts:
- Detect new patterns and trends. AI-based AML systems can be designed to detect previously unknown criminal behavior patterns, trends, and associations that are too complex to be picked up by straightforward, rule-based monitoring or the human eye. This could support the discovery of more convoluted money laundering schemes.
- Monitoring customer activity. One of the most important efficiencies AI can provide is the ability to analyze large amounts of data, filter out false alerts, and identify complex criminal conduct. AI can also detect unusual single transactions by creating an archetype of normal customer behavior and creating an alert when something out-of-the-ordinary arises.
- Compliance auditing. AI applications can support compliance in multiple ways, helping FIs overcome the challenges involved in a high volume of manual, repetitive, and data-intensive tasks. Automating compliance-based tasks allows for analysis of data in extremely high volumes and frequencies. Also, AI can reduce the number of false positive alerts, reduce operational workloads, and lower the high cost of compliance.
The future of AI in AML
Clearly, there is a great opportunity for the development and deployment of AI technologies to support our fight and enable the development of new AML strategies. In fact, AI’s potential could lead to a fundamental shift in AML operations.
However, along with the potential, the increasing awareness and number of AI applications has ignited some debate on the effectiveness of these solutions, and the extent to which AI can be trusted to replace human analysis and decision making. Among the concerns are algorithmic bias and lack of regulatory acceptance. What’s more, firms may find that they need to invest significantly—particularly at the outset— to ensure the competency of any systems, and in training employees to accurately monitor and interpret results.
To overcome the risks and realize the potential of AI, several actions need to take place. First, the AML community needs to continue building its understanding of the capabilities, risks, and limitations of AI. Secondly, we need to establish an ethical framework through which the development and use of AI can be governed, and the efficacy and impact of these emerging models can be proven and trusted.
While caution about wider AI adoption is easy to understand, it doesn’t have to slow down adoption of technologies that could usher AML into a whole new era. There is growing sentiment that by combining expert human insight and AI, we can drive better outcomes and new ways of working that are more effective than deploying either solution in isolation.
Fundamentally, this means that while AI will provide the brute force of powerful technology in financial crime fighting, there will always be a place for expert analysts to guide the technology in the right direction and manage the power of machine learning to supplement human expertise.