The Monetary Authority of Singapore (MAS) will soon finalise its AI risk management guidelines.
The guidelines will cover agentic AI alongside other AI use cases in financial institutions.
Published for consultation in November 2025, the proposed Guidelines on Artificial Intelligence Risk Management set out expectations for board and senior management oversight, risk management frameworks and processes, and controls across the AI life cycle.
MAS Chairman Gan Kim Yong said the regulator is taking a principles-based approach to support financial institutions in applying risk management practices proportionately as AI evolves.
The regulator has also worked with the industry on practical implementation tools.
Under Project MindForge, the industry developed an AI Risk Management Toolkit to help financial institutions apply the guidelines.
The Safeguards for Agentic Finance at Runtime (SAFR) framework sets out a potential approach for managing agent actions.
It covers how actions are authorised, when human oversight is activated and what is recorded when consequential decisions are made.
Gan was responding to the near-term risks posed by increasingly autonomous AI agents and whether SAFR could become a mandatory supervisory requirement.
MAS did not commit to making SAFR mandatory or provide a timeline for doing so.
Instead, the regulator said it will continue working with the industry through the Future of Finance Institute to develop good practices and toolkits while reviewing its supervisory expectations.
Featured image: Edited by Fintech News Singapore, based on image by Magnific
