Skip to main content

Free instrument from GAGE

Singapore governs AI without an AI law.

Ten instruments, four agencies, one map.

519 NODES · 824 EDGES · 27 INSTRUMENTS · 9 AGENCIES AND THE ASEAN RING · DATASET v2026.08.17 · VERIFIED 2026-08-17

The short answer

Singapore has no AI statute. Instead, four agencies, IMDA, PDPC, MAS and CSA, govern AI through ten instruments: binding data protection law and MAS notices, supervisory expectations for banks, consultation drafts, voluntary frameworks, one national standard, and security guidance. This map shows all of it, labelled by true legal force, with sources.

Legal force. Click a level to filter.

519 of 519 nodes shown

See it assemble, 2019 to 2026
Infocomm Media Development Authority (IM...Personal Data Protection Commission (PDPC)AI Verify Foundation (AIVF)Model AI Governance Framework, 2nd Editi...Guiding principles of the Model FrameworkPrinciple: explainable, transparent and ...Principle: human-centric AISector- and technology-agnostic designArea 1: Internal Governance Structures a...Adapt or set up internal governance stru...Define clear roles and responsibilitiesMaintain SOPs to monitor and manage AI r...Train staff on AI governancePeriodically review internal governance ...Area 2: Determining the Level of Human I...Assess probability and severity of harm ...Determine the appropriate degree of huma...Human-in-the-loopHuman-over-the-loopHuman-out-of-the-loopSeverity-probability of harm matrixArea 3: Operations ManagementMinimise bias in data and modelEnsure data qualityDocument data lineage and provenanceUse distinct datasets for training, test...Apply risk-based measures: explainabilit...Document model development choices and t...Monitor and regularly review deployed mo...Area 4: Stakeholder Interaction and Comm...Make AI policies known to usersAdopt a policy on explaining AI decisionsMake communications easy to understandAllow users to provide feedbackConsider offering opt-out where feasibleContinuously review communication effect...Annex A: Compilation of AI ethical princ...Annex B: Algorithmic accountability and ...Launch of Model AI Governance Framework,...Launch of Model AI Governance Framework,...Implementation and Self-Assessment Guide...ISAGO as companion guide to the Model Fr...ISAGO Section 1: Objectives of deploying...ISAGO Section 2: Internal governance str...ISAGO Section 3: Human involvement in AI...ISAGO Section 4: Operations managementISAGO Section 5: Stakeholder interaction...Define business objectives and weigh ben...Consider ethical implications of the AI ...Identify existing relevant governance st...Put in place an AI governance structure ...Self-assess the level of human involvementSelf-assess data management practicesSelf-assess model lifecycle practicesSelf-assess stakeholder communication pr...Launch of ISAGO at WEF DavosModel AI Governance Framework for Genera...Fostering a trusted AI ecosystemDimension 1: AccountabilityDimension 2: DataDimension 3: Trusted Development and Dep...Dimension 4: Incident ReportingDimension 5: Testing and AssuranceDimension 6: SecurityDimension 7: Content ProvenanceDimension 8: Safety and Alignment R&DDimension 9: AI for Public GoodAllocate responsibility ex-ante across t...Consider ex-post safety nets for end-usersClarify how personal data laws apply to ...Balance copyright with data accessibilityAdopt industry best practices in develop...Provide 'food label'-type transparency a...Establish incident monitoring and report...Use incidents for continuous improvementAdopt third-party testing and assuranceDevelop common AI testing standardsAdapt information security frameworks to...Develop new security testing toolsDeploy digital watermarking and cryptogr...Enable informed consumption of online co...Accelerate safety and alignment R&D inve...Cooperate globally on AI safety R&DDemocratise AI access and support public...Upskill the workforce and develop AI sus...Discussion Paper 'Generative AI: Implica...Proposed Model AI Governance Framework f...Finalised MGF for Generative AI releasedModel AI Governance Framework for Agenti...Dimension 1: Assess and bound the risks ...Dimension 2: Make humans meaningfully ac...Dimension 3: Implement technical control...Dimension 4: Enable end-user responsibil...Component: ModelComponent: InstructionsComponent: MemoryComponent: Planning and reasoningComponent: ToolsComponent: ProtocolsComponent: Controls (added in v1.5)Component: Logging and monitoring (added...Multi-agent pattern: SequentialMulti-agent pattern: SupervisorMulti-agent pattern: SwarmSystemic risks of multi-agent systems (v...Action-space (authority, capabilities)Autonomy (decision-making)Four levels of human involvement with ag...Five types of agentic AI riskAutomation biasAgent identity and access managementResponsibility allocation across the age...Computer use agentDetermine suitable use cases via agent-s...Perform threat modelling and taint tracingBound risks through agent limitsImplement agent identification and autho...Evaluate and accept residual riskClearly allocate responsibilities within...Practise adaptive governanceDefine significant human-approval checkp...Regularly audit the effectiveness of hum...Monitor human override rates and respons...Implement technical controls during desi...Test agents for baseline safety and reli...Roll out gradually and monitor continuou...Maintain change management for evolving ...Inform users of agent actions, data acce...Train users to manage human-agent intera...Maintain tradecraft and foundational ski...Launch of MGF for Agentic AI v1.0 at WEF...MGF for Agentic AI v1.5 updatePersonal Data Protection Act 2012 (No. 2...Part 3, General Rules on Protection of a...Part 4, Collection, Use and Disclosure o...Part 5, Access to and Correction of Pers...Part 6, Care of Personal DataPart 6A, Notification of Data BreachesPart 9, Do Not Call RegistryPart 9C, EnforcementConsent Obligation (s 13)Limits on Obtaining Consent (s 14)Deemed Consent (ss 15, 15A)Withdrawal of Consent (s 16)Collection, Use and Disclosure Without C...Purpose Limitation Obligation (s 18)Notification Obligation (s 20)Access Obligation (s 21)Correction Obligation (s 22)Accuracy Obligation (s 23)Protection Obligation (s 24)Retention Limitation Obligation (s 25)Transfer Limitation Obligation (s 26)Accountability Obligation (s 11)Designated Individual / DPO Requirement ...Policies and Practices Obligation (s 12)Duty to Assess Data Breaches (s 26C)Duty to Notify Notifiable Data Breaches ...Do-Not-Call Compliance (Part 9)'Personal Data' (s 2(1))'Data Intermediary' (s 2(1); s 4(2))'Publicly Available' Personal Data (s 2(...'Organisation' (s 2(1); s 4)'Notifiable Data Breach' (s 26B)Financial Penalties (s 48J)Directions for Non-Compliance (s 48I)PDPA Enacted (2012); Main Obligations in...Personal Data Protection (Amendment) Act...Breach Notification Regime in Force (1 F...Enhanced Financial Penalties in Force (1...PDPC Advisory Guidelines on Use of Perso...Part II, Legal Effect and ScopePart III, Using Personal Data in AI Syst...Part IV, Deployment: Collection and Use ...Part V, Procurement of AI Systems (B2B S...Meaningful Consent as the Default Basis ...Business Improvement Exception (as Appli...Conditions for Relying on the Business I...Using the Business Improvement Exception...Research Exception (as Applied to AI R&D)Conditions for Relying on the Research E...Legitimate Interests Exception (Deployme...Data Minimisation in AI Development (Goo...Data Protection Controls in the AI Devel...Anonymisation of Datasets for AI Develop...Notification Content for AI System Deplo...Accountability and Written Policies for ...Service Providers as Data IntermediariesData Mapping, Labelling and Provenance R...Using AI Verify to Support PDPA Complian...Public Consultation on Proposed Guidelin...Guidelines Issued (1 Mar 2024)PDPC Advisory Guidelines on Use of Perso...Public Consultation Opened (2 Jun 2026)Public Consultation Closed (1 Jul 2026; ...Final Guidelines Published (20 Jul 2026,...Development, Collecting and Using Person...Deployment, Data Protection Responsibili...Post-Deployment, Addressing Individuals'...Publicly Available Exception for GenAI T...'Digital Barrier' (Definition)Documented Assessment (DPIA or Written R...AI-Specific Notifications for GenAI Trai...General Notifications Insufficient for G...Consent for GenAI Training Cannot Be a C...Fresh Consent Where AI Use Departs Signi...Anonymisation Accepted as Alternative to...'User Data' (Personal Data From Products...Generative AI Stakeholders: Model Provid...Model Provider ResponsibilitiesSystem Provider ResponsibilitiesSystem Deployer Primary ResponsibilityAccess and Correction Obligations Apply ...Agentic Functionality Heightens Data Pro...AI Verify Testing Framework and ToolkitAI Verify Toolkit (Open Source)Technical Tests (Explainability, Robustn...Process Checks and Documentary EvidenceInternational Alignment of the Testing F...1. Transparency2. Explainability3. Repeatability / Reproducibility4. Safety5. Security6. Robustness7. Fairness8. Data Governance9. Accountability10. Human Agency and Oversight11. Inclusive Growth, Societal and Envir...12. Organisational ConsiderationsAI Verify MVP Launched (25 May 2022)AI Verify Open-Sourced on GitHub (7 Jun ...Crosswalk to NIST AI RMF Published (Oct ...Crosswalk to ISO/IEC 42001 Published (Ju...Enhanced Testing Framework for GenAI (29...Project Moonshot, LLM Evaluation ToolkitMoonshot (Software)Benchmarking ('Exam Questions' for LLMs)Red Teaming (Adversarial Prompting)Starter Kit for LLM-Based App TestingProject Moonshot Launched in Open Beta (...Global AI Assurance Pilot (Feb 2025)Pilot Launched at Paris AI Action Summit...Tester-Deployer Pairing ModelPilot Insights: Context-Dependent GenAI ...Pilot Converted to Global AI Assurance S...Monetary Authority of Singapore (MAS)Association of Banks in Singapore (ABS)Financial services sector (Singapore)Principles to Promote Fairness, Ethics, ...FEAT principle: FairnessFEAT principle: EthicsFEAT principle: AccountabilityFEAT principle: TransparencyFEAT Principle 01: Fairness - justifiabi...FEAT Principle 02: Fairness - justifiabi...FEAT Principle 03: Fairness - accuracy a...FEAT Principle 04: Fairness - accuracy a...FEAT Principle 05: Ethics)FEAT Principle 06: Ethics)FEAT Principle 07: Accountability - inte...FEAT Principle 08: Accountability - inte...FEAT Principle 09: Accountability - inte...FEAT Principle 10: Accountability - exte...FEAT Principle 11: Accountability - exte...FEAT Principle 12: Transparency)FEAT Principle 13: Transparency)FEAT Principle 14: Transparency)Definition of AIDAFEAT Principles published (12 Nov 2018)Veritas Initiative (MAS-led consortium)Veritas Phase 1: fairness assessment met...Veritas Phase 2: full FEAT assessment me...Veritas Phase 3: Toolkit v2.0 and integr...Veritas Toolkit (open source, v2.0)FEAT Fairness assessment methodologyFEAT Ethics & Accountability assessment ...FEAT Transparency assessment methodologyVeritas use case: credit risk scoringVeritas use case: customer marketingVeritas use case: insurance predictive u...Veritas use case: fraud detectionConsultation Paper on Proposed Guideline...P017-2025 consultation opened (13 Nov 20...P017-2025 consultation closed (31 Jan 20...Applicability and proportionality (Consu...Proposed Guidelines Section 2: AI oversi...Proposed Guidelines Section 3: key AI ri...Proposed Guidelines Section 4: key AI li...Proposed Guidelines Section 5: AI capabi...Annex: proportionate application of the ...Proposed scope of AI: model, system, use...Risk materiality dimensions: impact, com...Proposed 12-month transition periodProposed: board senior mgmt oversightProposed: dedicated ai committeeProposed: basic ai policiesProposed: ai identificationProposed: ai inventoryProposed: risk materiality assessmentProposed: data managementProposed: transparency explainabilityProposed: fairness controlsProposed: human oversightProposed: third party ai managementProposed: selection evaluation testingProposed: technology cybersecurityProposed: reproducibility auditabilityProposed: pre deployment reviewsProposed: post deployment monitoringProposed: incident management kill switchProposed: change management decommission...Proposed: capabilities trainingProposed: technology infrastructureMindForge AI Risk Management Toolkit (Pr...AI Risk Management: Executive Handbook (...AI Risk Management: Operationalisation H...AI Risk Management: Implementation Examp...Handbook Section 1: Scope and AI oversightHandbook Section 2: AI risk managementHandbook Section 3: AI lifecycle managem...Handbook Section 4: EnablersHandbook practice: define oversight resp...Handbook practice: ai policies standardsHandbook practice: org level risk manage...Handbook practice: third party ai riskHandbook practice: usecase level risk ma...Handbook practice: ai inventory capabili...Handbook practice: data acquisition proc...Handbook practice: onboarding build reviewHandbook practice: deployment controlsHandbook practice: usage monitoring changeHandbook practice: skills cultureHandbook practice: manage ai infrastruct...AI Card template (Handbook Appendix E)Library of AI metrics (Handbook Appendix...MindForge AI risk taxonomy (Handbook App...Project MindForge phase 2 concluded; too...MAS to establish AI risk management work...ABS Handbook on Generative AI Guardrails...Handbook Section 02: use case and risk i...Handbook Section 03: guardrails designHandbook Section 04: applying guardrails...Use case studies: document extraction/su...Guardrail approach: Enterprise Governanc...Guardrail approach: Filtering and ControlGuardrail approach: Customised Model Des...Guardrail approach: Red TeamingGuardrail approach: Prompt DesignGuardrail approach: Monitoring and Valid...Guardrail approach: Human-in-the-Loop Mo...Guardrail approach: User Feedback and It...Guardrail approach: User Transparency an...Seven enterprise Gen AI use case categor...Guardrails and Controls Excel toolHandbook published May 2025; current edi...MAS Notices on Technology Risk Managemen...Notice FSM-N05 Technology Risk Managemen...Notice FSM-N21 Technology Risk Managemen...Notice FSM-N13 Technology Risk Managemen...Notice FSM-N03 Technology Risk Managemen...TRM notice family across FI classesIdentify critical systemsMaximum 4 hours unscheduled downtime per...Recovery time objective of 4 hours or le...1-hour incident notification to MAS (the...Root cause and impact analysis report wi...IT controls to protect customer informat...Definition: relevant incidentDefinition: critical systemIncident notification instructions and r...Circular on Financial Institution Incide...Consultation P012-2026: proposed amendme...Cyber Security Agency of SingaporeEnterprise Singapore / Singapore Standar...ASEAN SecretariatNational Assembly of VietnamCSA Guidelines and Companion Guide on Se...Launch at SICW 2024 (15 October 2024)Public consultation (31 July - 15 Septem...Secure by design and secure by defaultTake a lifecycle approachStart with a risk assessmentStage 1: Planning and DesignStage 2: DevelopmentStage 3: DeploymentStage 4: Operations and MaintenanceStage 5: End of LifeStep 1: Conduct risk assessment focused ...Step 2: Prioritise areas to addressStep 3: Identify and implement relevant ...Step 4: Evaluate residual risks for miti...Raise awareness and competency on securi...Conduct security risk assessmentsSecure the supply chainConsider security benefits and trade-off...Identify, track and protect AI-related a...Secure the AI development environmentSecure the deployment infrastructure and...Establish incident management proceduresRelease AI systems responsiblyMonitor AI system inputsMonitor AI system outputs and behaviourAdopt a secure-by-design approach to upd...Establish a vulnerability disclosure pro...Ensure proper data and model disposalCompanion Guide on Securing AI SystemsThreat scope: supply chain attacks and A...Securing Agentic AI - Addendum to the Gu...Public consultation (22 October - 31 Dec...Official publication of Addendum v1.0 (1...Rogue actionsSensitive data disclosure through agent ...Three layers of agentic AI security riskBaseline components of agentic AI systemsComponent: Large Language ModelComponent: InstructionsComponent: ToolsComponent: MemoryComponent: ProtocolsCapability class: CognitiveCapability class: InteractionCapability class: OperationalAutonomy levels of agentic AI systemsAssess the autonomy level of the systemPerform threat modelling of agentic work...Apply taint tracing to track untrusted d...Identify risks associated with the agent...Step 1: Conduct a risk assessment focuse...Step 2: Prioritise areas to addressStep 3: Identify and implement relevant ...Step 4: Evaluate residual risks; re-eval...Agentic controls: Planning and DesignAgentic controls: DevelopmentAgentic controls: DeploymentAgentic controls: Operations and Mainten...SaaS agentic AI and shared responsibilityUse case examplesDefinition: agentic AI systemsSS ISO/IEC 42001:2024 - Information tech...Annex ZA (national, informative): AI Ver...Voluntary nature of Singapore StandardsPublication of SS ISO/IEC 42001:2024Scope: AI management system requirementsASEAN Guide on AI Governance and Ethics ...Endorsement at 4th ADGMIN (February 2024)Principle 1: Transparency and Explainabi...Principle 2: Fairness and EquityPrinciple 3: Security and SafetyPrinciple 4: Human-centricityPrinciple 5: Privacy and Data GovernancePrinciple 6: Accountability and IntegrityPrinciple 7: Robustness and ReliabilityComponent: Internal governance structure...Component: Determining the level of huma...Component: Operations managementComponent: Stakeholder interaction and c...Annex A: AI Risk Impact Assessment Templ...National-level recommendationsRegional-level recommendationsUse cases (Annex B)Expanded ASEAN Guide on AI Governance an...Launch at 5th ADGMIN (January 2025)Focus area 1: AccountabilityFocus area 2: DataFocus area 3: Trusted Development and De...Focus area 4: Incident ReportingFocus area 5: Testing and AssuranceFocus area 6: SecurityFocus area 7: Content ProvenanceFocus area 8: Safety and Alignment Resea...Focus area 9: AI for Public GoodSix GenAI risk categoriesUse casesASEAN Responsible AI Roadmap (2025-2030)Adoption of the Roadmap (5 March 2025)Focus area: Policy and regulatory founda...Focus area: Targeted actions, initiative...Readiness Assessment FrameworkASEAN Digital Economy Framework Agreemen...DEFA negotiations concluded (29 May 2026)Expected signature: 49th ASEAN Summit, N...Scope: emerging areas including AIVietnam Law on Artificial Intelligence N...National Assembly passage (10 December 2...Entry into force (1 March 2026)Risk-based classification: high, medium,...High-risk AI: pre-deployment assessment,...Transparency and AI-generated content la...Foreign providers: local presence or rep...Violations and liability (Article 29)Prohibited AI practicesRelationship to Digital Technology Indus...ISO/IEC 42001:2023 - AI management syste...Clause 4: Context of the organizationClause 5: LeadershipClause 6: PlanningClause 7: SupportClause 8: OperationClause 9: Performance evaluationClause 10: ImprovementNIST AI Risk Management Framework 1.0 (e...Function: GOVERNFunction: MAPFunction: MEASUREFunction: MANAGENote: NIST AI 600-1 Generative AI Profil...Regulation (EU) 2024/1689 - EU Artificia...Article 4: AI literacyArticle 6 + Annex III: High-risk AI syst...Article 9: Risk management systemArticle 10: Data and data governanceArticle 12: Record-keepingArticle 14: Human oversightArticle 15: Accuracy, robustness and cyb...Article 17: Quality management systemArticle 26: Obligations of deployersArticle 72: Post-market monitoring

Click the map to enable scroll zoom

Drag to pan. Click, then scroll to zoom. Double click to dive. Click a node to pin it, Escape to clear. Dashed amber edge means GAGE analysis, not an official mapping.

Which instruments can actually bind me?

61 entries on this map are binding law that applies to everyone in scope, and 17 bind MAS regulated financial institutions only. Everything else is a supervisory expectation, a consultation draft, a voluntary framework or a standard.

Who issues what

AgencyGuidance

Infocomm Media Development Authority (IMDA)

Singapore's Infocomm Media Development Authority; leads development of Singapore's AI governance instruments, including the Model AI Governance Framework family (2019-2026) and the world's first Model AI Governance Framework for Agentic AI.

AgencyBinding

Personal Data Protection Commission (PDPC)

Singapore's data protection authority. Administers and enforces the PDPA 2012, the only binding general law in Singapore's AI governance stack; issues advisory guidelines interpreting how PDPA obligations apply to AI systems.

AgencyGuidance

AI Verify Foundation (AIVF)

Not-for-profit foundation launched by IMDA in June 2023 to steward the open-source AI Verify testing framework and toolkit, grow the global AI testing community, and run international assurance initiatives.

AgencyBinding, sectoral

Monetary Authority of Singapore (MAS)

Singapore's central bank and integrated financial regulator. Issues binding notices to financial institutions under the Financial Services and Markets Act 2022 and other Acts, plus non-binding guidelines, information papers and consultation papers. legalForce value reflects that MAS issues BINDING-SECTORAL instruments; the actor node itself carries no obligation.

AgencyGuidance

Association of Banks in Singapore (ABS)

Industry association representing banks in Singapore. Its Standing Committee on Data Management (SCDM) published the Handbook on Generative AI Guardrails in Banking. ABS is an industry body, not a regulator; its publications are non-binding industry guidance. legalForce=GUIDANCE reflects that character.

AgencyGuidance

Cyber Security Agency of Singapore

Singapore's national cybersecurity agency, part of the Prime Minister's Office and managed by the Ministry of Digital Development and Information. Issued the Guidelines and Companion Guide on Securing AI Systems (2024) and the Addendum on Securing Agentic AI Systems (2026).

AgencyStandard

Enterprise Singapore / Singapore Standards Council

Singapore's national standards body administering Singapore Standards. Published SS ISO/IEC 42001:2024, the identical national adoption of ISO/IEC 42001:2023, prepared under the Information Technology Standards Committee.

AgencyGuidance

ASEAN Secretariat

Secretariat of the Association of Southeast Asian Nations (ten member states, eleven including Timor-Leste as observer). Hosts the ASEAN AI governance instruments: the 2024 Guide, the 2025 Generative AI expanded edition, the Responsible AI Roadmap 2025-2030, and DEFA negotiations.

AgencyBinding

National Assembly of Vietnam

Vietnam's legislature. Passed the Law on Artificial Intelligence No. 134/2025/QH15 on 10 December 2025 (15th National Assembly, 10th Session), in force 1 March 2026 - the first binding AI statute in ASEAN.

Start with the question you came in with

Common questions

Does Singapore have an AI law?
No. Singapore has no AI statute. Four agencies, IMDA, PDPC, MAS and CSA, govern AI through ten instruments: binding data protection law and MAS notices, supervisory expectations for banks, consultation drafts, voluntary frameworks, one national standard and security guidance.
Which Singapore AI instruments are actually legally binding?
The Personal Data Protection Act 2012 binds every organisation handling personal data, and MAS Technology Risk Management Notices bind MAS regulated financial institutions. The Model AI Governance Framework family, the CSA guidelines and AI Verify are voluntary. MAS Consultation Paper P017-2025 is proposed and not final.
Is the Model AI Governance Framework mandatory in Singapore?
No. The Model AI Governance Framework, including the Generative AI and Agentic AI editions, is voluntary guidance issued by IMDA. It becomes an obligation only where a contract, a sector regulator or a procurement requirement cites it.

The legend, in one line each

  • Binding61
  • Binding, sectoral17
  • Supervisory expectation21
  • Consultation44
  • Guidance358
  • Standard14
  • Emerging4

The same seven levels appear on every page, every zoom level and every badge. The full taxonomy is on the About page.

The Singapore AI Governance Navigator. GAGE is not affiliated with, endorsed by or acting for IMDA, PDPC, MAS, CSA, Enterprise Singapore, the AI Verify Foundation or any other body named on this map. Every entry links to the agency's own published source. Nothing here is legal advice.