Every AI Policy Ever · open in the interactive register

Published (non-binding, current)Non-binding · Guidelines / framework

Model AI Governance Framework for Generative AI

Record SG-04

Issued by
Infocomm Media Development Authority (IMDA) and AI Verify Foundation
Jurisdiction
Singapore
Level · region
National · South & Southeast Asia
Published / adopted
2024-05-30
Language
English

Summary

Framework for generative AI developed by IMDA and the AI Verify Foundation, expanding the Model AI Governance Framework for traditional AI. It sets out nine dimensions to be looked at in totality: accountability along the AI development chain, data quality and contentious training data, trusted development and deployment with 'food label'-type transparency, incident reporting, third-party testing and assurance, security against new threat vectors, content provenance such as watermarking, safety and alignment R&D, and AI for the public good. It calls on policymakers, industry, the research community and the public to do their part.

Safety topics

general-purpose / frontier AIincident reportingevaluation & testingtransparency & labellingaccountability & liabilityresearch priorities

Status notes

IMDA and the AI Verify Foundation released a proposed framework for international views on 16 Jan 2024, announced at Davos. They received over 70 responses from local and international companies, technology multinationals, audit firms and government agencies, and finalised the framework on 30 May 2024, when IMDA also announced that Singapore would lead, with Rwanda, the development of an AI Governance Playbook for the Digital Forum of Small States. The framework builds on the June 2023 Discussion Paper on Generative AI by IMDA, Aicadium and the AI Verify Foundation, and is meant as a first step: implementation guidelines and resources are to follow under each of the nine dimensions.

Sources

Open the official source https://www.imda.gov.sg/resources/press-releases-factsheets-and-speeches/factsheets/2024/gen-ai-and-digital-foss-ai-governance-playbook

Secondary source: https://aiverifyfoundation.sg/resources/mgf-gen-ai/

Local copy

Archived. Last attempt 2026-10-18 06:48:33 UTC

Official source (third-party copy, not published) archive/SG-04/2026-10-11/official.html
Retrieved 2026-10-11 14:02:25 UTC
SHA-256 2e3b60ff0859981dcf70d1be947f222a4b3a1f33fecd57a79eaec78af990497b
Secondary source (third-party copy, not published) archive/SG-04/2026-10-18/secondary.html
Retrieved 2026-10-18 00:48:36 UTC
SHA-256 92233b748ab792b4c4b5a37f4875eb4f7c91c822ab9cbece08009e16c5178eb6

Verification

B
Another official source, such as an announcement by the issuer, confirms the title, date and status. checked 2026-10-18 · alternative-official-url-fetched
Recheck of 18 Oct 2026. Stored and read: IMDA's factsheet of 30 May 2024 (imda.gov.sg, copy of 11 Oct 2026, unchanged when checked on 18 Oct 2026), whose text layer now includes its body: the proposed framework of January 2024, the more than 70 responses, the statement that the framework 'is finalised today' and the Digital FOSS playbook; and the AI Verify Foundation's framework page (aiverifyfoundation.sg, copy of 18 Oct 2026), which gives the full title, the nine dimensions and the release dates of 16 Jan 2024 and 30 May 2024. The framework's own text is published on aiverifyfoundation.sg, which is listed as not official, so it counts as secondary; the factsheet is an announcement, so the title, date and status rest on another official source (level B). The framework's PDF on aiverifyfoundation.sg (stored in this run; its file name says 19 June 2024, its text layer carries no date) gives the text behind the summary, the June 2023 Discussion Paper and the guidelines to follow, and the IMDA and AI Verify Foundation copyright (2024).
Second check: Independent check by Claude Code (claude-sonnet-5-5), 2026-10-18: 7 field(s) supported.

History

  1. 2026-10-18related: [{"slug": "singapore-model-ai-governance-framework-2020", "id": "SG-01"}, {"slug → [{"slug": "singapore-model-ai-governance-framework-2020", "id": "SG-01"}, {"slug — two-way link added: SG-06, SG-08 link(s) to SG-042026-10-18-fix-related-links · Claude Code (claude-opus-5-5)
  2. 2026-10-18verification.level: A → B — level computed from the evidence by tools/pipeline.py2026-10-18-refresh-level-rule-1 · Claude Code (claude-opus-5-5)
  3. 2026-10-18notes: IMDA and the AI Verify Foundation released a proposed framework for internationa → IMDA and the AI Verify Foundation released a proposed framework for internationa — the factsheet's body, now in the text layer, gives the more than 70 responses and the kind of respondents, and that the playbook for small states is developed with Rwanda; added. The 2023 Discussion Paper and the guidelines to follow rest on the framework PDF, stored in this run2026-10-18-refresh-level-rule-1 · Claude Code (claude-opus-5-5)
  4. 2026-10-18notes: Proposed as a draft in Jan 2024 (consultation to 15 Mar 2024; 70+ responses) and → IMDA and the AI Verify Foundation released a proposed framework for internationa — consultation deadline and response count (version 1) rest on no stored file and were removed; the remark about what the fetched factsheet showed moved out of the notes; the 2023 Discussion Paper and the planned implementation guidelines added from the framework PDF2026-10-18-recheck-south-asia-1 · Claude Code (claude-opus-5-5)
  5. 2026-10-18safety_topics: ["general-purpose / frontier AI", "incident reporting", "evaluation & testing", → ["general-purpose / frontier AI", "incident reporting", "evaluation & testing", — reduced from seven to the limit of six: deepfakes & synthetic media dropped; content provenance (watermarking, cryptographic provenance) is covered by transparency & labelling2026-10-18-recheck-south-asia-1 · Claude Code (claude-opus-5-5)
  6. 2026-10-18summary: Voluntary framework for generative AI. It sets out nine dimensions to be conside → Framework for generative AI developed by IMDA and the AI Verify Foundation, expa — rewritten from the framework PDF stored in this run and the AI Verify Foundation page: the nine dimensions as the framework describes them, its expansion of the traditional-AI framework and its call on all stakeholders2026-10-18-recheck-south-asia-1 · Claude Code (claude-opus-5-5)

Related records

Singapore Model AI Governance Framework Singapore · Published (non-binding, current)

AI Verify Singapore · Published (non-binding, current)

ASEAN Generative AI Guide ASEAN · Published (non-binding, current)

CSA Guidelines on Securing AI Systems Singapore · Published (non-binding, current)

Singapore MGF for Agentic AI Singapore · Published (non-binding, current)

Citation

Infocomm Media Development Authority (IMDA) and AI Verify Foundation. Model AI Governance Framework for Generative AI. 2024-05-30. https://www.imda.gov.sg/resources/press-releases-factsheets-and-speeches/factsheets/2024/gen-ai-and-digital-foss-ai-governance-playbook. Status: Published (non-binding, current); verified 2026-10-18 (level B). Local copy archive/SG-04/2026-10-11/official.html retrieved 2026-10-11 14:02:25 UTC, SHA-256 2e3b60ff0859981dcf70d1be947f222a4b3a1f33fecd57a79eaec78af990497b. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record SG-04.

Improve this record

Know of an amendment, a status change or an error? Report news or an error on GitHub; the form opens with this record filled in.