A new tool designed to help financial institutions detect money laundering has been launched by Google Cloud. Powered by artificial intelligence, the Anti Money Laundering AI (AML AI) service has already been tested by HSBC to monitor for suspicious activity.
Described as one of the biggest and most costly challenges facing the financial services sector, money laundering is an increasingly complex and global problem. According to Google, last year $2trn was laundered last year alone, with an average of between 2%-5% of global GDP pouring through the system annually.
This money is connected to activities like drug and human trafficking or funding terrorist activities. Tracking the money requires significant resources including hours of human analysis and work across multiple states, with reports to several different regulatory bodies.
To help tackle this problem Google has placed a big bet on machine learning tools. The tech giant says its AML AI service examines billions of records and transactions for trends and signs of financial crime. This, says Google, minimises wasted investigator time as it reduces the number of alerts and provides explainable outputs to speed up investigations.
As well as better tracking an explainability than existing offerings, Google says deploying AI also allows for more auditable and explainable outputs that can be used to defend decisions.
Legacy monitoring products rely on manually defined rules written from experience. This often results in high alert rates but low yields in terms of accuracy. Human-crafted rules are also easier for money launderers to learn and circumvent. Google suggests 95% of system-generated alerts from legacy tools are false positives. This is costly as it requires manual reviews that take time and distract from examining true suspicious activity.
Risk score for money laundering
AML AI creates a consolidated risk score as an alternative to a ruled-base alert. Rather than triggering a warning when certain pre-defined conditions are met, it tracks trends and behaviour to give a “money laundering score”. It is based on bank data including patterns, network behaviour and know your customer information. This lets the AI specifically monitor high-risk retail and commercial customers and also adapt to changes more quickly.
It is available alongside other Google Cloud AI products including the foundation model Vertex AI platform and data-heavy BigQuery tools. “Building on our commitment to bring AI-powered innovation to the financial services industry, we are launching Google Cloud’s AML AI to help financial institutions more accurately and efficiently identify AML risk while enhancing business operations and governance.” said Thomas Kurian, Google Cloud CEO.
HSBC was the test customer for the new product, finding it outperformed current systems in detecting financial crime risk. They found a two to four times increase in true positive risk and a 60% drop in alert volumes. This reduced operating costs and sped up detection.
“Google Cloud’s AML AI has significantly improved HSBC’s AML detection capability. Google’s models are already demonstrating the tremendous potential of machine learning to transform anti-financial crime efforts in the industry at large,” said Jennifer Calvery, group head of financial crime risk and compliance at HSBC.
“By enhancing our customer monitoring framework with Google Cloud’s sophisticated AI-based product, we have been able to improve the precision of our financial crime detection and reduce alert volumes meaning less investigation time is spent chasing false leads. We have also reduced the processing time required to analyse billions of transactions across millions of accounts from several weeks to a few days.”
This isn’t the first financial protection tool offered by Google Cloud. Last week, it put up a $1m crypto-mining protect programme for customers. The financial support was put in place because Google is confident it will not end up out of pocket thanks to the capabilities of its Security Command Center Premium tool, which offers specialised detection capabilities. These are engineered into the Google Cloud infrastructure.