AI in action · Regulation
Singapore's AI Rulebook for Finance Sets Two Deadlines
Singapore's central bank has finalised AI risk rules for every financial institution it supervises, AI agents included. Firms have until October 2027 and October 2028 to comply.
The Monetary Authority of Singapore published its final Guidelines on Artificial Intelligence Risk Management on October 7, 2026, almost a year after consulting on a draft in November 2025. They apply to all financial institutions MAS supervises, from banks, insurers, fund managers and financial advisers to exchanges, clearing houses, payment firms, custodians and credit rating agencies, and they cover every form of AI, including generative AI and AI agents.
The text is organised in four areas: oversight by the board and senior management; AI risk management systems, policies and procedures; controls over the AI life cycle; and the firm's own AI capability and capacity. The approach is risk-based rather than a list of bans. Firms are expected to identify where they use AI, assess how important and how risky each use is, and apply controls in proportion, scaled to their size, activities and exposure. MAS describes the guidelines as "high-level expectations that all FIs should adhere to."
Compliance comes in two stages. Governance and the risk management framework, the first two areas, should be in place by October 7, 2027; life-cycle controls and capabilities by October 7, 2028, according to summaries of the text by Baker McKenzie and Finance Magnates. The guidelines are supervisory expectations rather than legislation, and they sit alongside MAS's existing FEAT principles on fairness, ethics, accountability and transparency.
Agents get specific attention. MAS warns that AI agents with access to tools could take erroneous or unauthorised actions on their own, and singles out security, privacy and third-party risks for generative AI, on top of the financial, operational, conduct, financial-crime and reputational risks any AI system carries. The regulator plans to consult the industry in 2027 on whether further guidance on agentic AI is needed, Finance Magnates reported.
For customers the effect is indirect. A broker, fund manager or bank in Singapore that uses AI to recommend products, screen transactions or run an agent will have to know where each system sits, how material it is and who at the top answers for it. The guidelines arrived a week before the Financial Stability Board's final sound practices on AI adoption, published on October 15, and both point the same way: supervisors have stopped asking whether firms use AI and started asking them to show that they control it.
Sources: Monetary Authority of Singapore — Guidelines on Artificial Intelligence Risk Management (Oct 7, 2026); Finance Magnates — Singapore sets new AI risk rules for finance. What they mean for brokers (Oct 12, 2026); Baker McKenzie — Singapore: MAS issues Guidelines on Artificial Intelligence Risk Management (Oct 20, 2026)