Despite banks committing huge resources to the problem, the amount of money laundered each year is estimated to be two and five per cent of global GDP, or up to $2 trillion annually, according to the UN.

Google says that most legacy AML monitoring products are reliant on manually defined rules, which yield low rates of identifying suspicious activities. In fact, more than 95% of system-generated alerts turn out to be "false positives" in the first phase of review, with 98% never culminating in a suspicious activity report (SAR).

Google Cloud’s answer provides a consolidated machine learning-generated customer risk score as an alternative to rules-based transaction alerting. The risk score is based on the bank’s data including transactional patterns, network behaviour, and KYC data to identify instances and groups of high-risk retail and commercial customers.

The firm claims that its AML AI tech not only increases risk detection but also lowers operational costs and improves customer experience.

HSBC has been using the system, reporting that it has helped identify between wo and four times more suspicious activity, while reducing alert volumes by more than 60%.

"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," says Jennifer Calvery, group head, financial crime risk and compliance, HSBC.