What happened
In a written parliamentary reply dated for the August 5, 2026 Parliament sitting, MAS Chairman and Deputy Prime Minister Gan Kim Yong confirmed that the forthcoming Guidelines on AI Risk Management (AIRG) — first put out for consultation in November 2025 — apply to all AI use cases by MAS-regulated financial institutions, explicitly including agentic AI. MAS said it will maintain a principles-based (not prescriptive) approach, will finalize the guidelines soon, and is separately building implementation infrastructure through Project MindForge and the Safeguards for Agentic Finance at Runtime (SAFR) framework.
Why it matters
This makes MAS the first major financial regulator globally to explicitly confirm that autonomous AI agents fall within a binding (not merely voluntary) supervisory framework, ahead of the US Federal Reserve (which has excluded agentic AI from recent model risk guidance) and the EU AI Act (which has no agentic-specific category and delays high-risk enforcement to Dec 2027). It sets a global precedent that multinational financial institutions will need to track, and signals that 'principles-based' does not mean light-touch — non-compliance can trigger supervisory consequences even absent a specific rule violation.
Action needed
MAS-regulated banks, insurers, capital markets firms, and payment/FinTech licensees should begin compliance planning now for the AIRG guidelines' finalization (expected soon, timeline not yet fixed), focusing on board/senior management oversight, AI lifecycle governance (including third-party/vendor AI tools), and runtime governance/agent inventory for agentic systems.