The Monetary Authority of Singapore issued the Guidelines on Artificial Intelligence Risk Management on 7 October 2026, together with its response to feedback on the public consultation it ran in November 2025. The Guidelines are in final form and state supervisory expectations for financial institutions rather than statutory obligations.
Section 2 governs application: the Guidelines apply to all financial institutions and to all forms of AI technology, and each institution decides how to meet the expectations according to its own use of AI and the materiality of the risk. The substantive expectations sit in Section 3 on AI Oversight, Section 4 on Key AI Risk Management Systems, Policies and Procedures, Section 5 on AI Life Cycle Controls and Section 6 on AI Capability and Capacity. The Guidelines take effect on 7 October 2027. Implementation may be phased, with the expectations in Sections 3 to 4 applying from 7 October 2027 and those in Sections 5 and 6 applying by 7 October 2028.
Banks, insurers, capital markets services licensees and payment institutions supervised by MAS fall within scope, and so does AI obtained from third parties. The board and senior management are expected to own AI risk and to set roles, accountabilities and risk appetite, and MAS does not require a dedicated AI committee where existing governance structures can carry the work. Institutions are expected to maintain AI inventories, assess the materiality of each use, and apply proportionate controls covering data governance, testing, human oversight, cybersecurity, monitoring and change management. An institution stays accountable for AI supplied by a vendor, should obtain assurance over it and assess its suitability, and where the risk exceeds its risk appetite should consider limiting, suspending or replacing the service.
Where weak AI performance or unavailability is unlikely to have a material impact, MAS accepts that basic policies may suffice, and the depth of controls is expected to match risk exposure. MAS has said it intends to consult the industry in 2027 on additional guidance for agentic AI, so institutions deploying autonomous agents before that point work against expectations not yet published. The two compliance dates are 7 October 2027 and 7 October 2028.
Licentium advises banks, insurers, payment institutions, asset managers and technology vendors on Singapore financial regulation and on AI governance duties across jurisdictions. Work we undertake includes mapping current AI governance against Sections 3 to 6 of the Guidelines, AI inventory and risk materiality methodologies, third party AI assurance and contract terms, and phased implementation planning towards the 7 October 2027 and 7 October 2028 dates.
Source: MAS, Guidelines on Artificial Intelligence Risk Management, 7 October 2026