MAS sets expectations for how banks and insurers manage AI risk

Guidelines, not law, taking effect from 7 October 2027, in full by 7 October 2028.

MAS has issued Guidelines on Artificial Intelligence (AI) Risk Management, with a response to the feedback it received. They apply to all financial institutions and to all forms of AI technology, and set out supervisory expectations in four parts: strengthen oversight of AI risks with clear accountabilities; identify, assess and manage AI risks across the AI life cycle; manage the risks from third-party AI use; and apply the Guidelines in a risk-proportionate manner. Financial institutions can implement them in phases.

What this means for a Singapore SME

The Guidelines are addressed to financial institutions - a business's bank, its insurers and its insurance broker - not to the business itself. They reach an SME in two ways.

A company that sells AI tools or services to a bank, an insurer or another financial institution supplies third-party AI, so its product falls within the risks the Guidelines expect that institution to manage.

A company whose quote, claim or credit decision passes through an AI system at a financial institution deals with a firm that is expected to hold clear accountability for that system. The year before the Guidelines take effect is the window in which institutions build that oversight.