Model AI Governance Framework for Generative AI (2024)
Is this legally binding?
Guidance. Voluntary guidance. Best practice, not obligation, until a contract or a regulator cites it.
Finalised 30 May 2024 by AI Verify Foundation and IMDA after a proposed draft (16 Jan 2024) drew over 70 international responses; expands the Model Framework for traditional AI into nine dimensions to foster a trusted generative AI ecosystem. Non-binding guidance with practical suggestions for developers, deployers and policymakers.
From the source
“There are nine dimensions which the Framework proposes to be looked at in totality, to foster a trusted ecosystem.”
Executive Summary, pp.3-5; May 2024
Instrument record
- Short name
- MGF-GenAI
- Date
- 2024-05-30
- Version
- May 2024
- Cluster
- core
What this connects to
36 relations. Official relations are the ones the source documents state; anything marked GAGE analysis is our reading, not an agency's.
Contains9
- SectionDimension 1: AccountabilityGuidance
- SectionDimension 2: DataGuidance
- SectionDimension 3: Trusted Development and DeploymentGuidance
- SectionDimension 4: Incident ReportingGuidance
- SectionDimension 5: Testing and AssuranceGuidance
- SectionDimension 6: SecurityGuidance
- SectionDimension 7: Content ProvenanceGuidance
- SectionDimension 8: Safety and Alignment R&DGuidance
- SectionDimension 9: AI for Public GoodGuidance
Operationalised by1
Issued by2
Applied to by3
- EventDiscussion Paper 'Generative AI: Implications for Trust and Governance'GuidanceGAGE analysis, not official
Milestone event for the instrument
- EventProposed Model AI Governance Framework for Generative AI released for consultationConsultationGAGE analysis, not official
Milestone event for the instrument
- EventFinalised MGF for Generative AI releasedGuidanceGAGE analysis, not official
Milestone event for the instrument
Cites2
- InstrumentModel AI Governance Framework, 2nd Edition (2020)Guidance
GenAI MGF exec summary: MGF released 2019, updated 2020; framework expands on the existing Model Framework for traditional AI
- External frameworkNIST AI Risk Management Framework 1.0 (extern crosswalk target)GuidanceGAGE analysis, not official
IMDA 16 Jan 2024 press release situates the framework as building on the Singapore-US IMDA-NIST crosswalk of national AI governance frameworks
Cited by1
- InstrumentModel AI Governance Framework for Agentic AIGuidanceGAGE analysis, not official
Agentic MGF exec summary: 'Building on our previous model governance frameworks'; GenAI MGF (2024) is the immediate predecessor in the family
Maps across to7
- External frameworkNote: NIST AI 600-1 Generative AI Profile (July 2024)GuidanceGAGE analysis, not official
Both are generative-AI-specific governance layers over a base framework
- External frameworkArticle 15: Accuracy, robustness and cybersecurityBindingGAGE analysis, not official
GenAI MGF's security dimension anticipates EU accuracy/robustness/cybersecurity requirements
- External frameworkClause 8: OperationStandardGAGE analysis, not official
GenAI MGF operational dimensions map to AIMS operational controls
- External frameworkArticle 4: AI literacyBindingGAGE analysis, not official
GenAI MGF's capacity-building emphasis parallels Article 4 AI literacy
- External frameworkArticle 12: Record-keepingBindingGAGE analysis, not official
GenAI MGF documentation practices anticipate Article 12 logging
- External frameworkFunction: MEASUREGuidanceGAGE analysis, not official
GenAI MGF testing and assurance dimension corresponds to MEASURE
- External frameworkArticle 26: Obligations of deployersBindingGAGE analysis, not official
GenAI MGF deployer-facing guidance anticipates deployer obligations
Mapped across from11
- InstrumentExpanded ASEAN Guide on AI Governance and Ethics - Generative AI (2025)GuidanceGAGE analysis, not official
ASEAN GenAI focus areas mirror Singapore's GenAI MGF dimensions; Singapore championed the regional guide
- SectionFocus area 1: AccountabilityGuidanceGAGE analysis, not official
Mirrors GenAI MGF accountability dimension
- SectionFocus area 2: DataGuidanceGAGE analysis, not official
Mirrors GenAI MGF data dimension
- SectionFocus area 3: Trusted Development and DeploymentGuidanceGAGE analysis, not official
Mirrors GenAI MGF trusted development and deployment dimension
- SectionFocus area 4: Incident ReportingGuidanceGAGE analysis, not official
Mirrors GenAI MGF incident reporting dimension
- SectionFocus area 5: Testing and AssuranceGuidanceGAGE analysis, not official
Mirrors GenAI MGF testing and assurance dimension
- SectionFocus area 6: SecurityGuidanceGAGE analysis, not official
Mirrors GenAI MGF security dimension
- SectionFocus area 7: Content ProvenanceGuidanceGAGE analysis, not official
Mirrors GenAI MGF content provenance dimension
- SectionFocus area 8: Safety and Alignment Research & DevelopmentGuidanceGAGE analysis, not official
Mirrors GenAI MGF safety and alignment R&D dimension
- SectionFocus area 9: AI for Public GoodGuidanceGAGE analysis, not official
Mirrors GenAI MGF AI-for-public-good dimension
- InstrumentNIST AI 600-1: Generative AI ProfileGuidanceUnited StatesGAGE analysis, not official
The two national generative AI frameworks, published months apart
Verified against the official source on 2026-08-17. GAGE is not affiliated with or endorsed by any agency named here, and nothing on this page is legal advice. How this is built and checked.