The Monetary Authority of Singapore (MAS) is reportedly working with the banking industry and a law enforcement agency to test AI models that draw on cross-bank and public-private data, Managing Director Chia Der Jiun said in a speech at the Global FinTech Fest.
From bank-level monitoring to system-level detection
The initiative addresses a structural limitation in fraud controls: scam proceeds often move through accounts held at several different banks, meaning no single institution can see the full transaction chain. Through the process of combining information across participating organisations, the test aims to identify suspicious relationships earlier in the process.
According to MAS, the objective is to detect potential fraud sooner, intervene faster, and reduce resulting losses. The regulator has not disclosed which banks are taking part, which datasets are being used, or the criteria that will determine whether the models move beyond testing into deployment.
Alongside this initiative, MAS continues to run PathFin.ai, a programme that matches financial institutions with validated AI solutions. The programme is intended, in part, to lower the cost and technical burden smaller institutions face when adopting AI tools.
Regulators target different points in the scam chain
Other financial authorities are applying AI and analytics to fraud detection, though at different stages of the process. The Hong Kong Monetary Authority has directed banks to apply AI and network analytics when monitoring authorised payment scams, including the identification of complex networks of suspicious and mule accounts. It has also tested multi-bank information analysis in collaboration with the banking sector and law enforcement.
In the UK, the Financial Conduct Authority uses ML and web scraping to identify potentially fraudulent websites. In Australia, the Australian Securities and Investments Commission (ASIC) focuses on disrupting the online infrastructure used to attract victims, coordinating the removal of 11.964 phishing and investment scam websites in 2025, a 90% increase on the previous year.
FINRA, the US self-regulatory body for broker-dealers, applies machine learning at a different layer, analysing multiple market events for signs of fraud and manipulation rather than retail payment scams.
MAS expects results from its test by the end of 2026. The findings are expected to indicate whether banks, police, and the regulator can share and act on data quickly enough to interrupt scam flows before funds are dispersed beyond recovery.
