Singapore AI Governance Framework
Singapore's voluntary model AI governance frameworks: the Model AI Governance Framework (2nd ed., 2020) on internal governance, human involvement by probability and severity of harm, data and model operations (lineage, bias, explainability, robustness, traceability) and stakeholder communication; the Model AI Governance Framework for Generative AI (2024, nine dimensions); and the Model AI Governance Framework for Agentic AI (v1.5, May 2026) on bounding agent risks, human accountability and approval checkpoints, technical controls, testing and monitoring, and end-user responsibility. Guidance, not law.
Singapore AI Governance Framework is a compliance framework from Singapore with 9 domains and 28 controls that map to 150 other frameworks. The largest domains are GenAI framework: nine dimensions – Singapore AI Governance Framework (8 controls), MAF: operations management (data, algorithm and model) – Singapore AI Governance Framework (4 controls), Agentic AI framework: assess and bound the risks upfront – Singapore AI Governance Framework (3 controls). Every control below carries what it requires and what an assessor expects to see.
Framework summaries on this platform are interpretations grounded in the published source, for educational and compliance planning purposes. They do not reproduce or replace the official standards. Refer to the authoritative source for the definitive text. Framework names and trademarks belong to their respective organisations.
Framework Domains (9)
Agentic AI framework: assess and bound the risks upfront – Singapore AI Governance Framework
| Code | Title |
|---|---|
| singapore-ai-governance-framework::AGT-2.1.1 | Select agent use cases by assessing impact and likelihood factors and threat modelling |
| singapore-ai-governance-framework::AGT-2.1.2 | Bound agents by design: least-privilege tools, SOP-constrained autonomy and contained impact |
| singapore-ai-governance-framework::AGT-2.1.2-ID | Give every agent a unique, accountable, catalogued identity and scoped, bounded authority |
Agentic AI framework: enable end-user responsibility – Singapore AI Governance Framework
| Code | Title |
|---|---|
| singapore-ai-governance-framework::AGT-2.4 | Tell users they face an agent and what it can do, and train integrating users to oversee it and keep their skills |
Agentic AI framework: make humans meaningfully accountable – Singapore AI Governance Framework
| Code | Title |
|---|---|
| singapore-ai-governance-framework::AGT-2.2.1 | Allocate accountability across the agent value chain, internally and in contracts |
| singapore-ai-governance-framework::AGT-2.2.2 | Require human approval at significant checkpoints and audit that oversight stays effective |
Agentic AI framework: technical controls and processes – Singapore AI Governance Framework
| Code | Title |
|---|---|
| singapore-ai-governance-framework::AGT-2.3.1 | Prefer deterministic system-level controls over prompt guardrails, with runtime controls |
| singapore-ai-governance-framework::AGT-2.3.2 | Test agents end to end before deployment, individually and as a system, repeatedly and realistically |
| singapore-ai-governance-framework::AGT-2.3.3 | Roll out gradually, monitor and log in layers with defined interventions, and manage change |
GenAI framework: nine dimensions – Singapore AI Governance Framework
| Code | Title |
|---|---|
| singapore-ai-governance-framework::GEN-1 | Allocate accountability across the generative AI chain upfront and provide safety nets |
| singapore-ai-governance-framework::GEN-2 | Use trusted, quality training data and handle personal data and copyright pragmatically |
| singapore-ai-governance-framework::GEN-3 | Apply baseline safety practices, evaluate models and publish food-label disclosures |
| singapore-ai-governance-framework::GEN-4 | Run vulnerability reporting and AI incident reporting with a materiality threshold |
| singapore-ai-governance-framework::GEN-5 | Use independent, standardised third-party testing and assurance |
| singapore-ai-governance-framework::GEN-6 | Adapt security-by-design to generative AI threats with input filters and forensics |
| singapore-ai-governance-framework::GEN-7 | Label AI-generated content with watermarking or cryptographic provenance where practical |
| singapore-ai-governance-framework::GEN-8 | Support safety and alignment research and AI for public good |
MAF: human involvement in AI-augmented decision-making – Singapore AI Governance Framework
| Code | Title |
|---|---|
| singapore-ai-governance-framework::MAF-3.14 | Choose human-in, over- or out-of-the-loop oversight from the probability and severity of harm |
| singapore-ai-governance-framework::MAF-3.9 | Weigh commercial objectives against AI risks, guided by corporate values, in a reviewed risk impact assessment |
MAF: internal governance structures and measures – Singapore AI Governance Framework
| Code | Title |
|---|---|
| singapore-ai-governance-framework::MAF-3.7(1) | Allocate clear, trained and resourced roles for ethical AI deployment |
| singapore-ai-governance-framework::MAF-3.7(2) | Run risk management and internal controls specific to the deployed model |
MAF: operations management (data, algorithm and model) – Singapore AI Governance Framework
| Code | Title |
|---|---|
| singapore-ai-governance-framework::MAF-3.23 | Practise data accountability: lineage, quality, bias, separate datasets and periodic review |
| singapore-ai-governance-framework::MAF-3.27 | Make models explainable, or document repeatability where explanation is not feasible |
| singapore-ai-governance-framework::MAF-3.32 | Test robustness and tune models regularly in conditions like production |
| singapore-ai-governance-framework::MAF-3.37 | Build traceability, reproducibility and auditability in proportion to risk |
MAF: stakeholder interaction and communication – Singapore AI Governance Framework
| Code | Title |
|---|---|
| singapore-ai-governance-framework::MAF-3.46 | Disclose AI use and set a policy for which explanations are given and when |
| singapore-ai-governance-framework::MAF-3.52 | Offer opt-out or recourse, feedback and human review of material AI decisions |
| singapore-ai-governance-framework::MAF-3.55 | Test interfaces, communicate plainly, set acceptable use policies and get information from AI vendors |
Your Compliance Coverage
If you comply with Singapore AI Governance Framework, you already cover:
ASEAN Guide on AI Governance and Ethics
29%
14 controls mapped
Compare →UK AI Regulation Framework
8%
4 controls mapped
Compare →OECD AI Principles
8%
4 controls mapped
Compare →+ 147 more: Japan AI Guidelines (8%), IEEE 7000 (8%)
See all 150 mapped frameworks ↓Maps to 150 other frameworks
Coverage is not the same as your position
This page shows what Singapore AI Governance Framework overlaps with in general. Where your organisation actually stands, against the standard you are going for and the certifications you already hold, is a different question. Same graph and the same recorded refutations, scoped to you rather than to a pair.
The Compliance Position Diagnostic, $5,000 fixed, ten business daysWhat is Singapore AI Governance Framework and who does it apply to?
Singapore AI Governance Framework is a compliance framework from Singapore with 9 domains and 28 controls. Singapore's voluntary model AI governance frameworks: the Model AI Governance Framework (2nd ed., 2020) on internal governance, human involvement by probability and severity of harm, data and model operations (lineage, bias, explainability, robustness, traceability) and stakeholder communication; the Model AI Governance Framework for Generative AI (2024, nine dimensions); and the Model AI Governance Framework for Agentic AI (v1.5, May 2026) on bounding agent risks, human accountability and approval checkpoints, technical controls, testing and monitoring, and end-user responsibility. Guidance, not law. It is used by organisations to establish and maintain compliance with industry standards and regulatory requirements.
What does Singapore AI Governance Framework actually require?
Singapore AI Governance Framework has 28 controls organised across 9 domains. The largest domains are GenAI framework: nine dimensions – Singapore AI Governance Framework (8 controls), MAF: operations management (data, algorithm and model) – Singapore AI Governance Framework (4 controls), Agentic AI framework: assess and bound the risks upfront – Singapore AI Governance Framework (3 controls). Each control defines specific requirements that organisations must implement to achieve compliance.
If I already comply with another framework, how much of Singapore AI Governance Framework do I already cover?
Singapore AI Governance Framework maps to 150 other compliance frameworks. The top mapping partners are ASEAN Guide on AI Governance and Ethics (29% coverage), UK AI Regulation Framework (8% coverage), OECD AI Principles (8% coverage). Use our comparison tool to explore control-level mappings between frameworks.
How do I implement Singapore AI Governance Framework?
Start your Singapore AI Governance Framework compliance journey by running a self-assessment on our platform to identify your current compliance posture. Our AI advisory can answer specific questions about Singapore AI Governance Framework requirements, and cross-framework mapping helps you leverage existing controls from other frameworks you may already comply with. Create a free account to access all 28 controls and track your progress.
Start Your Compliance Journey
Create a free account to run self-assessments, get AI advisory, and track your compliance progress across 1,120 frameworks.
Get Started Free →Free forever, no credit card required