MAS Issues Guidelines on AI Risk Management to Promote Responsible Innovation

Picture Credit: International Accounting Bulletin

The Monetary Authority of Singapore (MAS) has released a consultation paper proposing new Guidelines on Artificial Intelligence (AI) Risk Management, marking a major step toward ensuring the responsible and safe use of AI in the financial sector. The proposed framework outlines supervisory expectations for all financial institutions (FIs) in managing AI-related risks while enabling innovation.

The Guidelines define MAS’s expectations in four key areas:

• Oversight of AI Risk Management: Boards and senior management must take accountability for establishing governance structures, policies, and a risk-aware culture for AI use.

• Risk Management Systems and Procedures: FIs are expected to maintain comprehensive AI inventories and conduct risk materiality assessments, considering each system’s impact, complexity, and reliance.

• AI Life Cycle Controls: The framework calls for robust controls over data management, fairness, transparency, explainability, human oversight, and third-party risks, applied proportionately to each use case.

• Capabilities and Capacity: Institutions must ensure their internal skill sets and resources are sufficient to deploy and monitor AI responsibly.

  
  

Recognizing that AI applications differ across institutions, the proposed Guidelines adopt a proportionate, risk-based approach, suitable for both traditional and emerging technologies, including Generative AI and AI agents.

The consultation builds on MAS’s 2024 thematic review of AI use in key banks and extensive dialogue with industry players. Ms Ho Hern Shin, MAS Deputy Managing Director, said the Guidelines will “provide financial institutions with clear supervisory expectations to support them in leveraging AI responsibly,” adding that they aim to “enable responsible innovation by financial institutions that implement the relevant safeguards.”