MAS and Industry Set Out Safeguards for Autonomous AI Agents in Finance

The Monetary Authority of Singapore, together with banks and fintech partners, has released a white paper setting out an industry framework to keep autonomous artificial-intelligence agents in finance operating safely, securely and within defined risk limits. The paper responds to the rapid emergence of AI agents that can carry out financial tasks with limited human intervention, a shift that makes continuous human oversight impractical as deployment scales. It proposes runtime safeguards and governance measures intended to ensure such agents behave predictably, stay within authorised boundaries and can be monitored, controlled and, where necessary, halted. The initiative reflects Singapore's broader strategy of encouraging financial innovation while managing the associated operational, security and accountability risks, and complements the regulator's wider work on technology risk, cyber resilience and AI governance. By convening industry participants to agree common principles, MAS aims to give firms a practical reference point for adopting agentic AI responsibly rather than leaving each institution to devise its own approach. The framework signals growing supervisory attention to the risks posed by increasingly autonomous systems across banking, payments and asset management, an area regulators worldwide are beginning to scrutinise.