Singapore's MAS Publishes Industry Framework to Keep AI Agents in Finance in Check

The Monetary Authority of Singapore, together with major financial institutions and fintech firms, has published an industry white paper setting out safeguards for artificial intelligence agents operating in finance. Titled Safeguards for Agentic Finance at Runtime, the paper proposes an industry-developed framework enabling AI agents in financial services to carry out tasks safely, securely and reliably. It was produced under BuildFin.ai, the regulator's programme supporting responsible development and deployment of AI across the financial sector. The initiative responds to a practical problem: as AI agents take on more autonomous work at a pace that makes hands-on human supervision impractical, firms need real-time controls that keep agent behaviour within defined mandates, policies and risk limits. The framework covers areas including policy-bound execution, real-time validation, auditability and interoperability controls. The paper is part of a broader Singapore push on financial technology that also includes a Future of Finance Institute announced in June to accelerate industry adoption of AI and tokenisation, and a Global FinTech Hackcelerator whose 2026 challenges focus on AI applications in credit and fraud risk modelling, digital-native wealth management and SME risk management. Singapore's financial sector contributes around 14% of GDP and employs about 200,000 people.