By Dan Byrne
Nasdaq has announced a new automated system as part of their latest AML efforts, which they have said will use artificial intelligence replicate complex human decisions.
In a press release Wednesday, the Nasdaq Automated Investigator was hailed as the “first automated solution for investigating AML for retail and commercial banks and other financial institutions.”
The system is intended to deal with the volume of alerts for possibly fraudulent transactions which can be generated through current AML monitoring practices.
This could be up to 300,000 alerts a month. Institutions which have narrowed their focus to more worthwhile alerts – potentially missing criminal activity in the process – could still end up with 25,000 per month, Nasdaq said.
The automated system simplifies the timeframe on alert analysis and uses any alert data available to it.
After collating and analysing transaction data, it will provide a clear, auditable reasoning behind all alerts – big or small – in the space of several seconds, Nasdaq said.
Nasdaq highlighted the system’s AI capabilities as the core element in how this reasoning is provided. The multiple AI variants involved can analyse through the lens of human decision-making to come to conclusions about each transaction based on that.
The AI will learn as it works in conjunction with business rules to become more efficient going forward. This learning could be either supervised or unsupervised, Nasdaq advised.
“Our solution improves the manually intensive anti-financial crime process by fully automating all stages of the existing manual alert investigation process to deliver accurate, consistent decision-making at scale,” said Darren Innes, head of AML Technology at Sell-Side and Buy-Side Solutions within Nasdaq. “This pinpoints the highest-probability of nefarious activity for analysts to prioritise, and provides them a layer of demonstrable explainability that does not currently exist in financial crime investigations.”
Share this on:
Follow us on: