Generative AI (Gen AI) has significantly empowered banks to enhance their creative
capabilities, streamline processes, and explore innovative solutions across various
facets of their operations. However, with these advancements come inherent risks and
challenges that must be managed effectively. To address these concerns, members of the
The Association of Banks in Singapore (ABS) Standing Committee for AI & Data
Management (SCADM) have collaborated to identify essential guardrails aimed at
mitigating the specific risks associated with the deployment of Gen AI, published ABS
Handbook on Generative AI Guardrails in Banking “Handbook” and a
supplementary excel tool.
This Handbook serves as an initial guide, drawing on the collective experiences of ABS
SCADM members to date and will provide in-depth guidance on how banks can select
appropriate guardrails tailored to their unique circumstances, taking into account the
characteristics and risk levels associated with their use of Gen AI. Additionally, the
Handbook will articulate where and how specific controls can be applied throughout the
Gen AI system lifecycle, offering practical illustrations of a range of implementation
controls that banks can adopt. It is designed to evolve alongside advancements in Gen AI
and the emergence of associated risks.
As the landscape of Generative AI continues to evolve, the ABS SCADM remains committed
to supporting banks in navigating the complexities of this technology while ensuring
robust governance and risk management practices are in place. The insights and
frameworks outlined in the upcoming Handbook will empower banks to harness the full
potential of Gen AI responsibly and effectively.
Building on the ABS Handbook on Generative AI Guardrails in Banking, members of the
Association of Banks in Singapore (ABS) Standing Committee on AI & Data Management
(SCADM) have developed the "From Principle-based Governance to Validation: Companion
Guide to the ABS Generative AI Guardrails Handbook (“Companion Guide”)". The
Companion Guide translates the principles established in the Handbook into practical
common use cases, associated risks and approaches to validating AI solutions, providing
a useful reference point to financial institutions (FIs) in assessing whether AI
controls and guardrails are effective, reliable and aligned with applicable regulatory
expectations.
The Companion Guide also complements MAS’ Guidelines on Artificial Intelligence
Risk Management for Financial Institutions, issued in October 2026. Collectively, these
initiatives strengthen the foundations for the safe and responsible use of AI across the
financial sector.