This ground-breaking solution leverages artificial intelligence to empower financial institutions in their relentless pursuit of identifying and curbing money laundering activities, which are estimated to account for 2-5% of global GDP annually.
In a world where conventional rule-based systems often falter, Google Cloud’s AML AI offers the opportunity to augment existing processes, promising a surge in risk detection accuracy, substantial operational cost savings, elevated governance standards and an improved customer experience.
In the first instance, the use of AI for detecting money laundering is challenging because of concerns about (1) risk coverage, (2) regulatory compliance and (3) model governance. This has been overcome by Google by only using in-house client data and secure deployment.
Understanding how the technology works
Google offer the product as a cloud-based gateway, an API, that scores AML risk based on in-house client data. It is leveraged as an enhancement to existing AML pipelines and SME resources.
The client's pre-prepared transaction, account, and AML data is passed to the Google Cloud AML AI API for (1) detection and (2) risk scoring. This is then returned to analysts for review and investigation. The solution is based on a dynamic feedback loop, so the model is regularly updated, and risk scores improve.
Unlocking AML AI's Transformative Benefits
At its core, AML AI introduces a range of transformative advantages. One of its most impressive feats is its unparalleled ability to amplify risk detection. A major player in the financial sector, HSBC, has witnessed an impressive 200-400% increase in true positive risk identification when using the new technology, significantly strengthening their ability to thwart money laundering schemes.
Furthermore, there is scope for a remarkable reduction in operational costs. Through regular training of the model, it is possible to achieve significantly fewer false positives through use of the Google AML AI API, thus enabling investigators to focus on higher value tasks. According to HSBC it was able to eliminate 60% of false positives with AML AI.
Another standout feature is the technology’s dedication to improved governance. A core feature of the product is model explainability: all risk scores are based on disclosed rationale and a confidence score. By generating transparent and comprehensible outputs the system provides invaluable support for internal risk management whilst ensuring compliance with a diverse array of regulatory demands across various global jurisdictions.
Moreover, the boost in customer experience cannot be overstated; the notable decline in false positives translates to fewer compliance checks, thereby streamlining banking operations and cultivating a more engaging customer interaction.
Facing Challenges and Critique
Of course, no groundbreaking innovation is without its challenges. For example, AML AI's effectiveness hinges on the volume, quality and completeness of the client data it processes, underscoring the critical need for meticulous data curation.
Another aspect to consider is that regulatory authorities and financial crime departments generally require financial institutions to provide a transparent explanation of the thought process underpinning the construction of their AML compliance programmes. This includes the methodology employed for fine-tuning their alert systems.
In addition, those same authorities and departments have often held on to the traditional practice of repeatedly reviewing all trades. This results in paying as much attention to low-risks entities as high-risk entities, rather than following a risk-based approach that facilitates fast and accurate detection of the most high-risk trades and customers. It could be argued that in this regard, Google’s solution is particularly innovative and targets a long-standing desire to shift frameworks.
Moreover, a question which some critics have posed is whether existing AML tools on the market can be as effective as humans at identifying where risk lies. Can Google’s offering prove that it can be not just as effective but even more so at identifying risks?
Demonstrating the Ability to Overcome Concerns
Where existing tools tend to focus on alerts for specific transactions only, Google’s AML AI is also able to identify groups of high-risk scoring customers, assign them a machine learning-generated customer risk score and detail factors which led to the score. Factors can include transactional patterns, network behaviour and Know Your Customer (KYC) data.
In addition, Google has evidenced through delivering AML AI to HSBC that concerns with satisfying regulators can be eased by sufficient testing and validation of the new tool.
The Path Forward
In conclusion, Google Cloud's launch of AML AI signals a pivotal moment in the fight against global money laundering. By harnessing the potential of advanced machine learning it aims to challenge the limitations of traditional rule-based systems, ushering in a new era characterized by heightened risk detection, optimised operational costs, elevated governance benchmarks and an enriched customer journey. With a success story already emerging from HSBC as well as Brazil’s Banco Bradesco and Denmark’s Lunar digital bank, AML AI holds the promise of steering the financial landscape towards a more automated and secure future.
How Delta Capita can help
Delta Capita has a successful track record in scaling AML/KYC operations rapidly to deliver projects of the highest quality, in line with client policies and timeframes. Our trained experts cover all aspects of the client due diligence review and can be deployed on-site or at one of our managed service centres – either using Delta Capita’s proprietary technology accelerators or, if preferred, the client’s own technology and platform tools.
To find out more and speak to one of our experts, contact us today.
.png)
.jpg)
.jpg)
.png)