MAS Puts AI at the Centre of Scam Prevention

Key Takeaways

• MAS launches AI proof‑value project with five major Singapore banks.
• Real transaction data, not synthetic, powers collaborative scam‑detection models.
• Hashed account numbers protect privacy while enabling cross‑bank pattern analysis.
• Success could expand AI use to broader financial‑crime use cases.
• Singapore positions itself as leader in regulated AI‑driven fraud prevention.

Pulse Analysis

Singapore’s Monetary Authority of Singapore (MAS) is moving beyond policy statements to operationalize artificial intelligence in the fight against financial crime. By partnering with the Government Technology Agency, the Singapore Police Force and five leading banks, MAS has created a live sandbox where real transaction histories feed machine‑learning models. This collaborative approach tackles a core weakness of traditional fraud systems: criminals often operate across multiple institutions, leaving single‑bank detectors blind to the full picture. Leveraging authentic data, the proof‑of‑value aims to flag suspicious accounts and transactions before losses occur, turning AI from a reactive tool into a proactive shield.

Privacy and governance sit at the heart of the initiative. MAS requires that all account identifiers be hashed, ensuring that only the originating bank can re‑identify its customers. This design balances the need for rich, cross‑bank signals with stringent data‑protection standards, addressing public concerns while preserving analytical depth. The secure data‑sharing environment also incorporates audit trails and model‑explainability requirements, aligning with emerging global AI‑governance frameworks. By testing both algorithmic performance and operational safeguards, MAS is setting a template for regulated AI deployment that other jurisdictions can emulate.

If the pilot demonstrates measurable reductions in scam losses, MAS plans to scale the model, incorporate additional financial‑crime use cases such as money‑laundering detection, and invite more institutions to join the data pool. Such expansion could accelerate the adoption of collective‑intelligence fraud defenses worldwide, prompting banks to shift from siloed solutions to industry‑wide AI ecosystems. For the payments sector, this could translate into faster, more resilient transaction networks, reinforcing consumer trust and positioning Singapore as a benchmark for AI‑enabled regulatory innovation.