MAS issues AI risk management guidelines for financial sector

MAS plans to consult on additional guidance on agentic AI in 2027.

The Monetary Authority of Singapore (MAS) has issued guidelines on artificial intelligence (AI) risk management for responsible AI adoption in the financial sector.

The guidelines, which take effect on 7 October 2027, set out supervisory expectations for financial institutions in Singapore to manage risks arising from AI use whilst allowing them to tailor their approaches according to their risk profiles, MAS said in a statement on 7 October 2026.

Financial institutions are expected to meet the expectations set out in Sections 3 and 4 of the documents by 7 October 2027, which cover board and senior management oversight and AI risk management systems, policies, and procedures.

They are expected to meet sections 5 and 6 by 7 October 2028. The sections tackle AI use case life cycles, data management, transparency and explainability, fairness, regular review controls, third party AI management, testing, risks, and AI risk management capabilities.

MAS’ four key expectations include strengthening oversight of AI risks with clear accountabilities; and identifying, assessing, and managing AI risks across the AI life cycle, including in data governance,testing, human oversight, cybersecurity, and monitoring.

MAS also expects financial institutions to manage the risks from third-party AI use and apply the guidelines in a risk-proportionate manner.

Financial institutions are expected to manage AI risks at both the enterprise and individual use case levels, and build the capabilities needed for responsible AI use as adoption grows.

Financial institutions are allowed to assess how best to meet the supervisory expectations based on the nature and scale of its AI use, and the associated risks, MAS said.

MAS said that financial institutions must review their controls regularly as their AI use expands, such as greater use of agentic AI systems.

The guidelines apply to all financial institutions and all forms of AI technologies, MAS said.

MAS cited a November 2025 public consultation, where financial institutions sought clarity on whether they can use existing governance structures.

Financial institutions also sought clarity on how to manage risks from embedded AI and when basic AI governance policies and procedures are sufficient, MAS said.

In 2027, MAS said that it intends to further consult the financial sector on what additional guidance on agentic AI would be useful.