(8 October 2026 – Singapore) Singapore’s Monetary Authority (MAS) has issued its inaugural Guidelines on Artificial Intelligence Risk Management for Financial Institutions, requiring banks to seek an independent review of AI projects prior to implementation.
The guidelines acknowledge AI’s potential to improve performance across the sector while cautioning that it can also introduce greater uncertainty, unexpected behaviour, and harder-to-detect bias compared to simpler methods.
The MAS flags agentic and generative AI as carrying heightened risk, warning that their greater complexity gives rise to even more uncertainty and unpredictable behaviour.
Financial institutions are held accountable for all AI used in their services — including systems developed or operated by third parties — and are expected to obtain sufficient assurance from vendors, assess suitability for intended use, and apply compensating controls where gaps exist. Where risks cannot be brought within a firm’s risk appetite, the MAS says institutions should consider limiting, suspending, or replacing the third-party AI service.
Additional requirements include maintaining a comprehensive inventory of all AI in use, managing risks arising from third parties that themselves deploy AI, and developing contingency plans for high-risk AI applications to ensure business continuity in the event of failure or unexpected behaviour.