Singapore
At a glance
- Regulator / body
- The Monetary Authority of Singapore (MAS) for the financial sector; the Infocomm Media Development Authority (IMDA) and the Ministry of Digital Development and Information (MDDI) for frameworks and national strategy; the Personal Data Protection Commission (PDPC) for personal-data guidance; and an AI Safety Institute through the Digital Trust Centre (details unverified).
- Status
- Singapore relies on voluntary frameworks plus a new sector-specific supervisory layer: the MAS AI risk guidelines were issued on 2026-10-07 but are not yet effective.
- Effective dates
- MAS Guidelines on AI Risk Management issued 2026-10-07; effective 2027-10-07, with sections 5 and 6 by 2028-10-07. IMDA agentic-AI framework launched 2026-01-22 (version 1.5 published 2026-05-20). National AI Strategy updated 2026-05-20.
- Last updated
- Oct 10, 2026
Recent developments and what to watch
- 2026-10-07:MAS issued its Guidelines on AI Risk Management(not yet effective; effective 2027-10-07). The consultation paper of 2025-11-13is superseded.
- 2026-07-20 (reported):PDPC guidance on personal data in generative AI.
- 2026-05-20:Agentic AI framework version 1.5and the National AI Strategy update.
What to watch:the MAS 2027 consultation on agentic-AI guidance and any move toward stronger safeguards for agentic AI.
Key instruments
No horizontal AI statute has been reported in Singapore. The main instruments are:
- MAS Guidelines on AI Risk Managementfor financial institutions (issued 2026-10-07). These set supervisory expectations. They areissued but not yet effective: they take effect on 2027-10-07, with sections 5 and 6 following by 2028-10-07. An earlier wiki labelling of these as in force was wrong and has been corrected.
- Model AI Governance Framework for Agentic AI(IMDA; version 1.0 launched 2026-01-22, version 1.5 published 2026-05-20). Voluntary guidance.
- National AI Strategy update(2026-05-20). Ten refreshed priorities and a layered, sector-specific approach to AI risk management.
- PDPC Advisory Guidelines on the Use of Personal Data in Generative AI, reported issued about 2026-07-20 (date rests on two secondary sources and needs confirmation). Non-binding interpretation of the Personal Data Protection Act.
Scope: who and what is covered
The MAS guidelines apply to financial institutions supervised by MAS. The IMDA framework and the PDPC guidance are voluntary and apply across sectors. MAS intends a 2027 consultation on agentic-AI guidance for financial institutions.
Compliance and enforcement
Financial institutions have until 2027-10-07 to meet the MAS guidelines, with sections 5 and 6 by 2028-10-07, and should plan accordingly. Outside finance, compliance is largely voluntary, though the PDPC's guidance indicates how the Personal Data Protection Act is likely to be applied to generative AI.
Caveats.The PDPC guidance date needs confirmation because the PDPC page body did not render. A paywalled report (MLex, 2026-10-08) that Singapore is studying stronger safeguards for higher-risk agentic uses is a signal only and is unverified. Research assistance, not legal advice.
References and further sources
Sources and further reading(prepared from the Spanglett Legal AI governance wiki and the October 2026 research run):
- MAS: supervisory expectations on responsible AI adoption (primary)
- MAS: 2025 consultation on the AI risk guidelines (superseded)
- IMDA: new Model AI Governance Framework for Agentic AI
- IMDA: Agentic AI framework (PDF)
- MDDI: update to the National AI Strategy (factsheet)
- PDPC: Advisory Guidelines on personal data in generative AI (date unconfirmed)
This page is research assistance, not legal advice. Verify against the primary sources before relying on it.
On this page
Research assistance, not legal advice. Verify against primary sources before reliance.