AI Guide For Boards in Singapore

AI GUIDE FOR BOARDS IN SINGAPORE

Section III: Governance and Oversight

7.5 Establishing AI Governance Policies

• How far is the organisation prepared to rely on opaque or “black-box” models? • What trade-os between eciency, transparency and accountability are acceptable?

Eective AI governance requires more than high-level principles.

Boards should ensure management translates governance expectations into operational disciplines through three interconnected layers:

7.5.2 Rules

• Intent. • Rules. • Processes.

Rules establish the minimum conditions required before AI systems are deployed.

Boards should expect material AI initiatives to include: • A named business owner where accountability lies. • Clearly-dened business case and strategic purpose. • Risk classication assessments. • Documented testing and validation protocols. • Dened human oversight processes. • Escalation and incident response mechanisms.

Boards must take the lead on decisions that balance governance and risk.

7.5.1 Intent

Intent denes the organisation’s strategic philosophy, ethical boundaries and governance expectations for AI deployment. Boards should guide management on: • Which decisions must remain explainable? • When must humans retain override authority?

67

Made with FlippingBook Digital Proposal Creator