Singapore MAS Guidelines on Artificial Intelligence Risk Management (2026) 5.5-5.6: Controls for transparency and explainability proportionate to the use case and its materiality, with greater attention for credit, insurance, advisory and fund management uses to input features, key drivers, informing customers, consequences and redress, tailored to the audience

What does 5.5-5.6 require?

An FI establishes controls for an appropriate level of transparency (disclosures to those affected by AI use and explanations where relevant and requested) and explainability (methods that make AI outputs or decisions understandable, built in or post hoc) for each AI use case, so that customers, internal users and other stakeholders can judge the risks, reliability and limitations of the AI and the FI can show its use aligns with its stated objectives and standards. The degree required reflects the context, nature and assessed materiality of the use case: AI relied on heavily for credit decisioning, insurance underwriting, financial advisory or fund management needs more than AI assisting internal users in low-risk tasks. For higher-transparency uses the FI pays greater attention to the features used as inputs and the justification for them, users' ability to identify key drivers of the output, informing customers that AI is used, the consequences of AI-driven decisions and channels for redress, and tailors the manner of transparency to the audience (non-technical for customers, more technical for internal evaluators and overseers).