AI Guide For Boards in Singapore

AI GUIDE FOR BOARDS IN SINGAPORE

Section III: Governance and Oversight

A Risk-Tiered Approach to AI Governance

High Risk

Category 3: External-Facing / Customer-Impacting AI (Strict Controls, Continuous Monitoring, Human-in-the-loop)

Category 2: Internet Business Workflows & Operations (Process Optimisation, Bias Testing, Role Reskilling)

Low Risk

Category 1: Personal Productivity & General Use (Frictionless Enablement, Safe Data Guardrails)

7.1.1 Personal productivity

AI acceptable use policies, AI-specic data classication guidance, enterprise- managed AI environments, data loss prevention mechanisms, and workforce training on responsible AI usage. Boards should also ensure that existing enterprise risk and compliance frameworks are updated to address AI usage across the workforce. New and emerging forms of AI use, such as agentic AI and physical AI, may need additional governance, including in areas such as workplace safety. Board checklist for management • Are employees provided with AI tools and trained on the responsible use of AI tools? • Do enterprise policies clearly dene permitted and prohibited AI usage? • Are corporate data and

• Opportunities: General-purpose AI tools (such as meeting summarisation, content draing and code generation) are increasingly used to support individual productivity. ese applications may improve eciency and reduce administrative burden across the workforce. • Risks: e primary risks relate to unmanaged data exposure, inappropriate use of public AI tools, leakage of condential information and loss or misuse of personal data. Employees may inadvertently input proprietary source code, sensitive customer data or condential material into unmanaged or unapproved AI tools. Boards should distinguish these risks from enterprise- managed AI environments that are governed by contractual, administrative, access and data-protection controls. Use of unapproved AI tools can create data protection risks, intellectual property risks, cyber security exposures and regulatory compliance concerns. • Board considerations: Boards must ensure management establishes clear

intellectual property protected through enterprise-managed AI environments?

• Have existing risk and compliance frameworks been updated to address AI usage?

58

Made with FlippingBook Digital Proposal Creator