Model AI Governance Framework for Agentic AI
Is this legally binding?
Guidance. Voluntary guidance. Best practice, not obligation, until a contract or a regulator cites it.
World's first governance framework for agentic AI, launched by IMDA at WEF Davos on 22 January 2026; gives organisations a structured overview of agentic AI risks and emerging best practices across four dimensions. Version 1.0 (22 Jan 2026); v1.5 update released 20 May 2026 with industry feedback.
From the source
“The Model AI Governance Framework (MGF) for Agentic AI gives organisations a structured overview of the risks of agentic AI and emerging best practices in managing these risks.”
Cover; Executive Summary, pp.1-2
Instrument record
- Short name
- MGF-Agentic
- Date
- 2026-01-22
- Version
- v1.0; v1.5 update 2026-05-20
- Cluster
- core
What this connects to
30 relations. Official relations are the ones the source documents state; anything marked GAGE analysis is our reading, not an agency's.
Contains4
Operationalised by12
- ConceptComponent: ModelGuidance
- ConceptComponent: InstructionsGuidance
- ConceptComponent: MemoryGuidance
- ConceptComponent: Planning and reasoningGuidance
- ConceptComponent: ToolsGuidance
- ConceptComponent: ProtocolsGuidance
- ConceptComponent: Controls (added in v1.5)GuidanceGAGE analysis, not official
v1.5-added component; primary v1.5 PDF not retrievable at build time
- ConceptComponent: Logging and monitoring (added in v1.5)GuidanceGAGE analysis, not official
v1.5-added component; primary v1.5 PDF not retrievable at build time
- ConceptMulti-agent pattern: SequentialGuidance
- ConceptMulti-agent pattern: SupervisorGuidance
- ConceptMulti-agent pattern: SwarmGuidance
- ConceptSystemic risks of multi-agent systems (v1.5)GuidanceGAGE analysis, not official
v1.5 named systemic risks verified via secondary sources only
Applied to by2
- EventLaunch of MGF for Agentic AI v1.0 at WEF DavosGuidanceGAGE analysis, not official
Milestone event for the instrument
- EventMGF for Agentic AI v1.5 updateGuidanceGAGE analysis, not official
Milestone event for the instrument
Cites3
- InstrumentModel AI Governance Framework, 2nd Edition (2020)Guidance
Agentic MGF s.2: builds on the responsible AI practices set out in MGF (2020), footnote 14
- InstrumentModel AI Governance Framework for Generative AI (2024)GuidanceGAGE analysis, not official
Agentic MGF exec summary: 'Building on our previous model governance frameworks'; GenAI MGF (2024) is the immediate predecessor in the family
- InstrumentSecuring Agentic AI, Addendum to the Guidelines and Companion Guide on Securing AI SystemsGuidance
Agentic MGF footnotes 1, 12 and text adapt from CSA Draft Addendum on Securing Agentic AI
Cited by1
- InstrumentPDPC Advisory Guidelines on Use of Personal Data in Generative AI (final, 20 Jul 2026)Guidance
Para 9.6 refers deployers of agentic GenAI systems to IMDA's Model AI Governance Framework for Agentic AI
Maps across to5
- External frameworkArticle 9: Risk management systemBindingGAGE analysis, not official
Bounding agentic risks upfront anticipates the Article 9 risk management system
- External frameworkFunction: GOVERNGuidanceGAGE analysis, not official
Agentic governance dimensions extend GOVERN to autonomous systems
- External frameworkClause 6: PlanningStandardGAGE analysis, not official
Upfront risk bounding for agents aligns with AIMS risk planning
- External frameworkArticle 15: Accuracy, robustness and cybersecurityBindingGAGE analysis, not official
Agentic technical controls anticipate robustness and cybersecurity requirements
- External frameworkFunction: MANAGEGuidanceGAGE analysis, not official
Lifecycle technical controls for agents correspond to MANAGE risk treatment
Mapped across from2
- InstrumentMindForge AI Risk Management Toolkit (Project MindForge phase 2)GuidanceGAGE analysis, not official
Thematic crosswalk: Handbook extends AI risk management to agentic AI; IMDA's Model AI Governance Framework for Agentic AI is the national agentic framework. GAGE analysis, not an official mapping
- SectionThe AI Agent Standards InitiativeGuidanceUSGAGE analysis, not official
The two national agentic AI programmes: Singapore wrote a framework, the United States started a standards initiative
Learn this properly
This page tells you what Model AI Governance Framework for Agentic AI is and whether it binds you. The AI Governance program teaches the whole discipline, with dedicated coverage of the Singapore governance stack and the MAS regime, and every topic is passed by explaining it back in your own words, graded against the source.
See the AI Governance programVerified 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.
Readers of this also ask
GAGE briefings tell you which AI regulation deadlines are coming, what they actually require of you, and when a program opens.