Framework details
Framework name and version:
IMDA Model AI Governance Framework for Agentic AI, Version 1.5
Published 20 May 2026 (Updated 5 June 2026)
Framework URL (publicly accessible):
https://www.imda.gov.sg/-/media/imda/files/about/emerging-tech-and-research/artificial-intelligence/mgf-for-agentic-ai.pdf
Source list(s) to map:
Justification
Why does this framework matter for GenAI security practitioners?
It's the first governance framework built specifically around agent-native concepts, when agent identity distinct from human identity, considering action-space/autonomy as risk factors, and addressing multi-agent cascading effects. Not a general AI risk guidance adapted for AI agents. Developed by Singaporean regulator with input from government agencies and 60+ industry organisations.
Is this framework used in a regulated sector or jurisdiction?
Published by IMDA, Singapore's statutory infocomm/media regulator. Positioned as a best-practice guidance, but functions as de-facto national reference for agentic AI governance in Singapore. Already being referenced as a working standard in APAC.
What gap does this fill that existing mappings don't cover?
APAC-specific frameworks
Coverage estimate
How many of the source list entries map to this framework?
7–9 of the 10 Agentic Top 10 entries
Are you willing to write the mapping file(s)?
Conflict of interest
Are you affiliated with the framework body or a vendor that benefits from this mapping?
Not affiliated with IMDA.
Disclosure: I run an AI governance consultancy, so I have a professional interest in being represented, but no commercial relationship with IMDA itself.
Framework details
Framework name and version:
IMDA Model AI Governance Framework for Agentic AI, Version 1.5
Published 20 May 2026 (Updated 5 June 2026)
Framework URL (publicly accessible):
https://www.imda.gov.sg/-/media/imda/files/about/emerging-tech-and-research/artificial-intelligence/mgf-for-agentic-ai.pdf
Source list(s) to map:
Justification
Why does this framework matter for GenAI security practitioners?
It's the first governance framework built specifically around agent-native concepts, when agent identity distinct from human identity, considering action-space/autonomy as risk factors, and addressing multi-agent cascading effects. Not a general AI risk guidance adapted for AI agents. Developed by Singaporean regulator with input from government agencies and 60+ industry organisations.
Is this framework used in a regulated sector or jurisdiction?
Published by IMDA, Singapore's statutory infocomm/media regulator. Positioned as a best-practice guidance, but functions as de-facto national reference for agentic AI governance in Singapore. Already being referenced as a working standard in APAC.
What gap does this fill that existing mappings don't cover?
APAC-specific frameworks
Coverage estimate
How many of the source list entries map to this framework?
7–9 of the 10 Agentic Top 10 entries
Are you willing to write the mapping file(s)?
Conflict of interest
Are you affiliated with the framework body or a vendor that benefits from this mapping?
Not affiliated with IMDA.
Disclosure: I run an AI governance consultancy, so I have a professional interest in being represented, but no commercial relationship with IMDA itself.