AI Regulation Reference

Model AI Governance Framework

Jurisdiction:
Singapore
 voluntary
Effective:
Jan 23, 2019
Authority:
Infocomm Media Development Authority
 Official text Verified Mar 26, 2026

Obligations Covered

Human Oversight Explainability Risk Assessment

AI Governance and Human Oversight #

Obligation:
Human Oversight
 voluntary
Effective:
Jan 23, 2019
Risk tier:
all
Scope:
providers, deployers

Requirements

RequirementDetails
Human-in-the-loopAppropriate level of human involvement based on risk and impact
Decision modelsThree models: human-in-the-loop, human-on-the-loop, human-out-of-the-loop
Risk-proportionateLevel of oversight proportionate to risk of AI application
Agentic AI oversight2026 update adds guidance for autonomous agent monitoring and intervention

Penalties

ViolationFine
Non-complianceVoluntary — no direct penalties; organizations legally accountable for AI actions under existing law

Explainability and Transparency #

Obligation:
Explainability
 voluntary
Effective:
Jan 23, 2019
Risk tier:
all
Scope:
providers, deployers

Requirements

RequirementDetails
Explainable AIProvide explanations of AI decisions appropriate to the audience
TransparencyDisclose use of AI in decision-making to affected individuals
Stakeholder communicationProactive communication about AI use, capabilities, and limitations

Penalties

ViolationFine
Non-complianceVoluntary — no direct penalties

AI Risk Management and Third-Party Oversight #

Obligation:
Risk Assessment
 voluntary
Effective:
Jan 23, 2019
Risk tier:
all
Scope:
providers, deployers

Requirements

RequirementDetails
Internal governanceEstablish AI governance structures and accountability
Risk managementLifecycle risk management from design through deployment and monitoring
Third-party oversightAssess and manage risks from AI vendor and third-party systems
Agentic AI risks2026 update covers system design, deployment safeguards, monitoring, and end-user responsibility

Penalties

ViolationFine
Non-complianceVoluntary — no direct penalties