Singapore regulator wants independent review of financial firms’ AI systems

Singapore’s financial regulator wants firms to have every AI use case independently reviewed before it goes live. The guidance also calls for ongoing monitoring, including of third-party AI, and says firms remain accountable for AI used in the services they provide.

Watch Desk analysis

What happened

Singapore’s Monetary Authority (MAS) has published its first Guidelines on Artificial Intelligence Risk Management for Financial Institutions, The Register reports. The guidelines are due to come into force on 7 October 2027.

MAS calls for independent reviewers, not involved in developing an AI system, to check that controls such as evaluation and testing have been followed before deployment. It also wants cybersecurity and technology reviews, ongoing monitoring for performance changes, and up-to-date inventories of AI used by financial institutions.

Why it matters

The guidance addresses a practical problem for banks and other financial firms: an AI service may come from a supplier, but that does not make its risks someone else’s responsibility. MAS says firms should seek assurance from providers, assess whether a system suits its intended use and add controls where assurance is lacking. If risks cannot be brought within a firm’s tolerance, the regulator says it should consider limiting, suspending or replacing the service.

For high-risk uses, the guidance also calls for contingency plans, including alternative systems or manual processes. That is less glamorous than an AI launch, but considerably more useful when a system stops behaving as expected.

Our read

This is a substantial governance framework, not a blanket ban on using AI. Its most consequential message is that financial firms should be able to identify, review and monitor AI even when a supplier built it. The practical test will be whether those expectations lead to meaningful checks, rather than another tidy document filed beside the risk register.

What to watch

• How Singaporean financial firms implement the reviews before the guidelines take effect.
• What assurance firms can obtain from third-party AI providers, especially when their systems are opaque.
• Whether MAS clarifies how it expects firms to manage AI that suppliers use without disclosing it.

Discussion spark: Should financial firms be allowed to deploy third-party AI when the supplier cannot provide enough assurance, or should the firm’s accountability mean the answer is simply no?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.