IN THE ongoing fight against financial crime such as money laundering and terrorist financing, financial institutions must stay ahead of the new ways criminals are trying to infiltrate the banking system. Increasingly, technologies such as artificial intelligence (AI), machine learning (ML) and robotics process automation have become the cavalry deployed by banks to bolster their frontline defence.
As criminals become more sophisticated, tweaking their methods and leveraging emerging technologies to try to avoid detection, financial institutions must similarly arm themselves with new technologies to tackle these constantly evolving threats.
Investments into technologies and innovation for anti-money laundering (AML) and other financial crime compliance systems are becoming more and more critical for banks to keep abreast of these threats.
The pressure and demands on compliance capabilities will continue to mount, and it is important to be agile in ensuring preventative measures are effective, to ensure that the financial system does not become a conduit for illicit flows of funds.
AI SPEEDS UP DETECTION, ACTION, UPHOLDING TRUST
AML standards within Singapore's banking system are one of the highest and most robust globally. At the base level, all banks have in place AML systems that raise alerts on suspicious banking transactions. These are done through automated systems that perform checks through indicators (called risk dimensions) based on predetermined rules.
Integrating AI in these rules-based AML systems supercharges the fight against money laundering. The technology greatly speeds up the process of detecting high-priority suspicious transactions so that thorough investigations can take place earlier.
For example, UOB is the first Singapore bank to apply AI concurrently to two AML risk dimensions - transaction monitoring and name screening. Its AI-driven AML system today analyses 60,000 account names monthly against global regulatory watch lists. UOB can also pinpoint more accurately higher-priority cases from the more-than-5,700 average monthly suspicious transaction alerts flagged. This is since the models used for name screening and transaction monitoring have achieved 96 per cent prediction accuracy in the 'high priority' category.
In integrating AI in its AML system, UOB is able to be more deliberate and swifter in deploying resources to investigate potential money laundering attempts.
AI INVESTMENTS HINGE ON TRUST FROM STAKEHOLDERS
Implementing any AI solution, especially for regulated entities such as banks, requires multiple considerations. For instance, stakeholders and regulators must trust the use and results of the AI model and the algorithms that drive it. One way in which financial institutions can engender buy-in is through ensuring rigorous tests, which can validate the effectiveness and transparency of the AI model.
Instilling trust in stakeholders towards the use of AI requires clear mapping of how this emerging technology is used, what risks are involved and how they are managed. There should also be professional validation of the reliability of the AI model.
Hence, it is essential for financial institutions to come up with the relevant frameworks, perimeters and rubrics to convince stakeholders while driving implementation. As a case in point, when implementing AI-driven AML, UOB worked with Deloitte to develop an objective framework to guide key aspects of the model's risk governance and to test, assess and validate the AI model architecture. This journey was documented in a white paper titled "Advanced analytics and innovation in Financial Crime Compliance - The future is now".
The digitalisation of work and services brings about new dimensions of risks, and strengthening risk management capabilities must be a key focus for financial institutions. The results of AI and ML in AML systems, such as that which UOB has implemented, present a strong case for the adoption of these technologies to stay ahead of criminals and to safeguard customers and the financial system.
This journey would not be fruitful without collaborating as a wider ecosystem. When regulators and industry players come together to share practices, it will be a win-win as the industry moves forward to combat financial crime more effectively.
The next step is to further deepen collaboration and accelerate the building of capabilities to bolster trust and build an ecosystem, and create industry-level monitoring utilities incorporating AI and ML, thereby strengthen the industry's capabilities to take on sophisticated financial crimes.
It is important for financial institutions to start the journey now and to stay the course to continue to win the battle against financial crime.
- The writer is financial crime compliance leader at Deloitte South-east Asia