The paper proposes new Guidelines on Artificial Intelligence (AI) Risk Management and sets supervisory expectations for the responsible, risk-based use of AI in the financial sector.
Key insights include:
➡️ Oversight & Governance: The proposed Guideline set clear expectations for board and senior management accountability, requiring firms to establish governance, oversight, and policies for managing AI risks;
➡️ AI Risk Management: Financial institutions (FIs) would be required to maintain accurate AI inventories, and assess risk materiality based on system complexity, impact, and reliance;
➡️ Lifecycle Controls: Firms would be advised to apply proportionate controls for data, transparency, fairness, third-party risk, and human oversight across the AI lifecycle, ensuring capabilities match their AI exposure; and
➡️ Scope & Proportionality: The proposed Guidelines would apply to a broad range of AI, including GenAI and emerging agents.
✅ Firms should review existing AI governance frameworks, and strengthen oversight of model risk, explainability, and data governance ahead of the finalisation of MAS’s AI Risk Management Guidelines in 2026.