MAS Turns to AI to Fight Financial Crime
Singapore’s financial regulator is exploring a new approach to an old problem: stopping scams before they drain bank accounts. The Monetary Authority of Singapore (MAS) is collaborating with local banks, the Government Technology Agency, and the Singapore Police Force to apply machine learning techniques to fraud detection.
The initiative is currently structured as a proof-of-value (POV) that combines transaction data from five banks. The idea is straightforward: more data across institutions could help build more accurate AI models. If those models can flag high-risk transactions and accounts earlier, banks might be able to intervene before customers lose money.
Data sharing across banks is tricky territory when customer information is involved. MAS has set up a secure environment with specific safeguards. Bank account numbers are hashed, which means only the bank that contributed the data can reverse the conversion and identify actual account holders. Access is limited to authorized personnel within a monitored setting, and all data will be deleted when the POV concludes.
For now, this is a testing phase. MAS plans to evaluate the POV’s effectiveness before considering any expansion. If the approach proves valuable, future iterations could involve broader datasets and additional use cases, strengthening defences against criminal activity.
This work is part of a broader industry push to harness AI and machine learning for anti-financial crime efforts, complementing existing bank-level controls rather than replacing them.
