Singapore Sets AI Risk Rules for Financial Institutions from 7 October 2027

by admin
Singapore Sets AI Risk Rules for Financial Institutions from 7 October 2027

Singapore’s central bank has issued final guidelines on how all financial institutions should manage AI risks, taking effect on 7 October 2027.

The Monetary Authority of Singapore (MAS), the country’s central bank and financial regulator, on 7 October 2026 issued final guidelines that set how banks, insurers, payment firms and other financial institutions are expected to manage risks from their use of artificial intelligence, taking effect on 7 October 20271.

The Guidelines on Artificial Intelligence Risk Management for Financial Institutions cover all financial institutions and all forms of AI technology2.

In short, every Singapore financial firm that uses AI will need board-level oversight, a record of where it uses AI, and controls scaled to the risk. Firms whose AI use is low-stakes may keep basic policies, while heavier users face fuller controls over a two-year phase-in.

What do the MAS AI guidelines require?

MAS says the guidelines set clear supervisory expectations, meaning the standards its supervisors will check firms against1. Firms are expected to manage AI risks at two levels: across the whole organization and for each individual use of AI1.

From consultation to full compliance

Nov 2025 MAS consults on draft
7 Oct 2026 Final guidelines issued
7 Oct 2027 Guidelines take effect
7 Oct 2028 Phase-in ends

The release groups the main expectations under four headings:

  • Stronger oversight of AI risks, with clear lines of responsibility
  • Identifying, assessing and managing AI risks across the whole life of an AI system
  • Managing risks from AI supplied by outside vendors
  • Applying the rules in proportion to the risk

These expectations were kept from the draft MAS consulted on in November 2025, and refined to answer industry feedback1.

Who is responsible inside a financial institution?

The board and senior management should provide effective oversight of AI risks1. That includes setting roles and responsibilities, a risk appetite, which is the level of risk a firm is willing to accept, and risk management frameworks, policies and procedures1.

Firms may use their existing governance structures where these give adequate oversight and coordination across departments1. MAS says firms need not set up a dedicated AI committee only to meet this expectation1. This answers a question raised by respondents during the consultation1.

How must firms track and control their AI use?

Firms should identify where they use AI and keep an inventory of AI use at a suitable level of detail1. They must then assess how material the risk of each use case is, and apply controls in proportion1.

What firms are expected to do with their AI

Identify AI use Keep an inventory at a suitable level of detail
Assess risk materiality For each AI use case
Apply proportionate controls Data governance, testing, human oversight, cybersecurity, monitoring
Review controls regularly As AI use grows and technology changes

Those controls run through the full life of an AI system, from build to retirement1. They include data governance, testing, human oversight, cybersecurity, monitoring and change management1.

Firms must review their controls regularly as their AI use grows and the technology changes1. MAS points to agentic AI, systems that can act on their own and use other software tools, as one such change1.

What about AI bought from outside vendors?

Firms remain accountable for AI used in the services they deliver, even when a third party built, runs or supplies it1. This matters for banks and insurers that rely on vendor chatbots, scoring models or software with AI built in.

Firms should get enough assurance from vendors and judge whether a vendor’s AI suits the intended use1. Where gaps in assurance remain, they must add compensating controls, meaning extra safeguards of their own1. If the risk still cannot be kept within the firm’s risk appetite, the firm should consider limiting, suspending or replacing the vendor’s AI service1.

Which Singapore financial firms are covered?

The guidelines apply to a long list of licence types2. These include full, merchant and wholesale banks, finance companies and credit and charge card issuers2. They also cover fund managers, financial advisers, exchanges, clearing houses and credit rating agencies2.

Payment firms are in scope, including major and standard payment institutions and money-changing licensees2. So are life, general and composite insurers, reinsurers, captive insurers and insurance brokers2.

Smaller users get lighter treatment. A firm may rely on basic AI policies and procedures if poor performance or failure of its AI tools is unlikely to have a material impact on the firm, its customers or other stakeholders, including other financial firms1. The depth of controls should depend on each firm’s risk exposure1.

When do the AI rules take effect?

The guidelines take effect on 7 October 20271. Firms may phase in compliance1. They should meet the expectations in Sections 3 to 4 from 7 October 2027, and those in Sections 5 and 6 by 7 October 20281.

Two-stage phase-in for all financial institutions

DateWhat falls due
7 Oct 2027Guidelines take effect; Sections 3 to 4 expectations apply
7 Oct 2028Sections 5 and 6 expectations must be met

MAS has also published a response to feedback paper alongside the guidelines1. The Financial Stability Board, an international body that monitors the global financial system, has separately consulted on sound practices for financial firms’ use of AI1.

The next step is on agentic AI: in 2027, MAS intends to consult the financial sector again on what extra guidance on such systems would be useful1.

Sources

  1. ↩ MAS Sets Out Supervisory Expectations on Responsible AI Adoption by Financial Institutions https://www.mas.gov.sg/news/media-releases/2026/mas-sets-out-supervisory-expectations-on-responsible-ai-adoption-by-financial-institutions
  2. ↩ Guidelines on Artificial Intelligence Risk Management for Financial Institutions https://www.mas.gov.sg/regulation/guidelines/guidelines-on-artificial-intelligence-risk-management-for-financial-institutions

You may also like