The Monetary Authority of Singapore (MAS) has operationalised its approach to artificial intelligence (AI) governance in financial services, publishing a new AI Risk Management Toolkit developed through industry collaboration under Project MindForge. The initiative brings together 24 banks, insurers and capital markets firms, reflecting a coordinated effort to translate high-level AI principles into practical implementation frameworks.
At the centre of the release is an AI Risk Management Operationalisation Handbook, designed to provide firms with actionable guidance on embedding AI risk controls across both traditional and emerging use cases, including generative and agentic AI. The toolkit focuses on how institutions can implement and evidence existing risk management expectations in production environments.
The handbook is structured around four core areas aligned with MAS’ proposed AI risk management guidelines. These include governance and oversight, where firms are expected to define clear accountability for AI systems; risk management practices, including use-case identification and materiality assessment; lifecycle controls spanning development through deployment and monitoring; and a set of organisational enablers, such as data infrastructure and internal capabilities required to sustain responsible AI adoption.
Accompanying the handbook is a set of industry case studies, offering insight into how financial institutions are approaching AI deployment in practice. These examples highlight the operational challenges associated with scaling AI safely, particularly as firms move beyond experimental use cases into business-critical applications. They also reflect the growing need to integrate AI governance into existing risk and compliance frameworks rather than treating it as a standalone discipline.
The publication comes as MAS continues to review feedback on its earlier consultation on AI risk management guidelines, suggesting that supervisory expectations in this area are still evolving. In this context, the toolkit can be seen as an intermediary step—bridging policy intent and operational execution—while allowing both regulators and firms to iterate on best practices.
MAS has indicated that the handbook will be updated over time as industry adoption matures and supervisory expectations become more defined. This iterative approach recognises the pace of change in AI technologies, particularly with the emergence of more autonomous and agent-based systems that introduce new categories of model and conduct risk.
To support ongoing development, MAS will establish an industry workgroup under its BuildFin.ai initiative, bringing together MindForge participants and additional stakeholders. The group is expected to focus on developing implementation resources, sharing practical experience, and advancing approaches to managing risks associated with newer AI paradigms.
Commenting on the initiative, Kenneth Gay, Chief FinTech Officer at MAS, said: “The development of the MindForge AI Risk Management Toolkit, including the release of the Operationalisation Handbook, marks a major step forward in our journey to ensure the responsible adoption of AI in finance. We are committed to fostering a culture of continuous engagement and strengthening of AI governance and risk management practices across the industry. The BuildFin.ai programme also serves as a foundation for our next phase of collaboration in AI risk management, to bolster the safe adoption of AI across the financial industry.”
Taken together, the toolkit and associated industry collaboration point to a shift towards more collaborative, governance-led approaches to AI. As deployment extends across front-to-back workflows, the ability to operationalise risk controls, and evidence their effectiveness to supervisors become core conditions for deploying AI safely and credibly at scale.
Subscribe to our newsletter
