RT AI TOOLKIT

Best Practices for AI Governance and Risk Management Published for Singapore Financial Sector

In March 2026, the Monetary Authority of Singapore (“MAS“) and the Association of Banks in Singapore (“ABS“) released the following publications, respectively, aimed at guiding financial institutions in the responsible adoption of artificial intelligence (“AI“):

  1. The AI Risk Management Toolkit (“Toolkit“) released by MAS on 20 March 2026; and
  2. The Handbook on Generative AI Guardrails in Banking (“Handbook“) published by the Standing Committee on Data Management (“SCDM“) of ABS on 24 March 2026.

MAS AI Risk Management Toolkit

The Toolkit is released under phase two of Project MindForge and was developed with contributions from a range of industry participants, including banks, insurers and capital market firms. The Toolkit features an AI Risk Management Operationalisation Handbook (“Operationalisation Handbook“) containing practical guidance on implementing AI risk management framework, supported by AI case studies drawn from participating financial institutions.

MAS is currently reviewing feedback on the Proposed MAS Guidelines for Artificial Intelligence Risk Management which was the subject of a consultation exercise from 15 November 2025 to 31 January 2026. The Operationalisation Handbook is structured around the following four sections aligned with the proposed Guidelines:

  1. Scope and AI oversight: addresses the establishment of AI governance structures and the clarification of AI oversight roles within financial institutions;
  2. AI risk management: covers the identification of AI usage, risk materiality assessment and building AI inventory capabilities through appropriate systems, policies and procedures;
  3. AI lifecycle management: concerns the application of controls throughout the various stages of the AI lifecycle; and
  4. Enablers: focuses on building the organisational capabilities, infrastructure, and resources necessary for responsible AI use and ongoing risk management.

To encourage wider adoption of AI risk management practices, MAS will form an AI risk management workgroup with MindForge consortium members and other industry experts under BuildFin.ai. This workgroup will create implementation resources, support knowledge sharing and develop frameworks for managing risks arising from emerging AI technologies, including agentic AI.

ABS Handbook on Generative AI Guardrails in Banking

The Handbook provides a clear framework for financial institutions to adopt generative AI (“Gen AI“) in a responsible, secure and ethical manner, with a focus on strong governance, technical controls and human oversight.

In particular, the Handbook identifies guardrails, being pre-established guidelines and processes that act as a safety net to prevent and mitigate potential risks, to address the specific risks that may arise when deploying Gen AI. The approach taken was informed by the experience of SCDM members implementing seven common categories of enterprise Gen AI use across over 30 use cases in their organisations. This was followed by the identification of the most prevalent and applicable Gen AI risks, and corresponding guardrails to mitigate them.

The Handbook describes how banks should select the appropriate guardrails for their situation based on the characteristics and level of risk of their use of Gen AI. An accompanying Excel Tool describes these guardrails in detail, articulates where in the Gen AI system lifecycle they should be applied, and illustrates a range of specific implementation controls that a bank can adopt in practice. Practical considerations around these guardrails are further illustrated through two common use case studies on document extraction and summarisation and code generation. These use case studies describe the common risks and key considerations for implementing the relevant guardrails in a real-world context.

Click on the following links for more information:

      Available on the MAS website at www.mas.gov.sg: 

      Available on ABS website at www.abs.org.sg: