DatasetCC-BY 4.0 Open Datav2.1 (Updated Q1 2026)

Mapping responsible AI initiatives across Asia

A systematic, comparative regional inventory tracking 248 government roadmaps, binding audit directives, regulatory sandboxes, and civil society oversight initiatives across 14 Asian jurisdictions.

Scope: 14 Jurisdictions (South & Southeast Asia)

Records: 248 catalogued initiatives

Period: 2022 – 2026

TOTAL INVENTORY

248

Policy instruments

STATUTORY / BINDING

21.8% legally binding

VOLUNTARY / GUIDES

112

45.2% non-binding

CITIZEN AUDIT

19.0% redress routes

Direct Data Access & Codebook

Download the cleaned tabular data (CSV) or read the variable codebook defining classification schemas, operationalized metrics, and data curation ethics.

EXPLORE DATA REPOSITORY

Sample Records Explorer (10 records shown)

Filter by instrument classification or country to inspect regional governance variations.

ID	Country	Initiative & Lead Entity	Instrument Type	Status	Governance Scope
IND-01	India	Digital Personal Data Protection (DPDP) Algorithmic Processing Rules Ministry of Electronics & IT (MeitY) (2024–2026)	Statutory Law	Binding	Automated Decision Making & Consent: Statutory constraints on state and private algorithmic profiling, mandating verifiable consent protocols and exemptions review for biometric public welfare distribution.
IDN-02	Indonesia	SE Menkominfo No. 9/2023 on Artificial Intelligence Ethics Ministry of Communication & Informatics (Kominfo) (2023–2026)	Voluntary Guideline	Advisory	Ethical Principles & Algorithmic Accountability: National ethical baseline demanding transparency, audit trails, and human-in-the-loop oversight for state agencies deploying generative AI in municipal public services.
LKA-03	Sri Lanka	National AI Strategy & Public Procurement Safeguards Roadmap Ministry of Technology / ICTA (2024–2028)	Statutory Law	Binding	Public Procurement & Linguistic Equity: Sets strict pre-procurement algorithmic risk assessments for public administrative tools, with specialized criteria for Sinhala and Tamil language validation.
SGP-04	Singapore	Model AI Governance Framework for Generative AI & AI Verify Toolkit IMDA & Personal Data Protection Commission (2024–2026)	Regulatory Sandbox	Pilot	Testing Sandboxes & Transparency Benchmarks: Open-source governance testing suite providing 11 dimensions for red-teaming foundational models, hallucination mitigation, and copyright compliance.
PHL-05	Philippines	National AI Strategy Roadmap 2.0 & Algorithmic Ethics Registry Department of Trade and Industry (DTI) (2024–2026)	Voluntary Guideline	Advisory	Workforce Disruption & Public Algorithmic Registry: Establishes a centralized public registry for automated scoring models in social security, prioritizing worker protection and public sector automation transparency.
THA-06	Thailand	Royal Decree on AI System Classification & Clinical AI Sandbox Electronic Transactions Development Agency (ETDA) (2025–2026)	Regulatory Sandbox	Pilot	High-Risk Systems & Healthcare Sandboxes: Tiered risk taxonomy classifying public diagnostic AI as high-risk, requiring institutional sandbox approvals and independent clinical safety audits.
REG-07	Regional	Southeast Asia Responsible AI Monitor (SEARAM) Civil Society Consortium for Digital Rights (2023–2026)	Watchdog / Civil Society	Oversight	Surveillance Technologies & Citizen Redress: Cross-border grassroots observatory documenting facial recognition deployments, police automated dispatching, and community appeal bottlenecks.
BGD-08	Bangladesh	National Artificial Intelligence Policy & Civic Tech Framework ICT Division / Aspire to Innovate (a2i) (2024–2026)	Statutory Law	Binding	Welfare Delivery & Bengali NLP Equity: Mandates public service delivery bots to undergo bias testing for rural dialects and socio-economic marginalization indicators before public tender approval.
VNM-09	Vietnam	National AI Standards & Algorithmic Impact Assessment Guidelines Ministry of Science and Technology (MOST) (2024–2026)	Voluntary Guideline	Advisory	Critical Infrastructure & Smart Cities: Technical benchmarks for government smart city traffic monitoring and public sensor arrays, specifying data residency and model explainability obligations.
MYS-10	Malaysia	National AI Governance and Ethics Guidelines (AIGE) Ministry of Science, Technology and Innovation (MOSTI) (2024–2026)	Voluntary Guideline	Advisory	Data Sovereignty & Public Sector Procurement: Focuses on trustworthy algorithmic deployment across healthcare and education ministries, featuring specific clauses on preventing algorithmic bias against indigenous populations.

IndiaIND-01

Binding

Digital Personal Data Protection (DPDP) Algorithmic Processing Rules

Ministry of Electronics & IT (MeitY) • 2024–2026

Statutory Law

Automated Decision Making & Consent: Statutory constraints on state and private algorithmic profiling, mandating verifiable consent protocols and exemptions review for biometric public welfare distribution.

IndonesiaIDN-02

Advisory

SE Menkominfo No. 9/2023 on Artificial Intelligence Ethics

Ministry of Communication & Informatics (Kominfo) • 2023–2026

Voluntary Guideline

Ethical Principles & Algorithmic Accountability: National ethical baseline demanding transparency, audit trails, and human-in-the-loop oversight for state agencies deploying generative AI in municipal public services.

Sri LankaLKA-03

Binding

National AI Strategy & Public Procurement Safeguards Roadmap

Ministry of Technology / ICTA • 2024–2028

Statutory Law

Public Procurement & Linguistic Equity: Sets strict pre-procurement algorithmic risk assessments for public administrative tools, with specialized criteria for Sinhala and Tamil language validation.

Showing 3 of 10 initiatives

* Note: Showing representative sample subset from the 248 records. Download full CSV for raw schema including citation sources, gazette serials, and statutory review timelines.

METHODOLOGICAL PROTOCOL

Data collection and classification ontology

The Asia AI4D inventory systematically documents emerging state directives, national strategies, administrative decrees, and independent monitoring platforms across 14 jurisdictions. Primary data is extracted directly from official legislative gazettes, ministerial releases, parliamentary question-hour transcripts, and verified civil society litigation filings.

Each initiative is coded across five standardized ontological axes:

Enforceability Level: Distinguishing legally binding statutory instruments from non-binding ministerial circulars and voluntary industry codes.
Transparency & Algorithmic Registry Mandates: Evaluating whether public agencies are compelled to disclose algorithms used in citizen welfare allocation, policing, or tax audits.
Redress Mechanisms: Verification of whether citizen appeals are routed through independent ombudsman bodies or restricted to vendor dispute clauses.
Linguistic Representation: Assessing if evaluation protocols mandate testing across local official dialects and low-resource vernaculars.
Procurement Oversight: Tracking requirements for algorithmic impact assessments (AIA) prior to awarding public cloud or AI vendor tenders.

Key findings from the 2026 inventory

Analysis of the complete 248-record corpus highlights three prevailing regional trends:

Voluntary guidelines outpace statutory mandates by more than 2:1. While governments rapidly publish ethical frameworks, statutory accountability mechanisms with fiscal penalties or independent judicial review remain concentrated in only 4 jurisdictions.
Procurement remains the primary governance blindspot. Less than 24% of tracked instruments mandate algorithmic impact assessments before public agencies procure proprietary vendor solutions, creating structural vendor lock-in.
Emergence of civic oversight networks. Civil society watchdogs represent the fastest-growing sector of initiatives (+42% YoY), actively filling institutional audit gaps in social protection and municipal surveillance deployments.

← Back to knowledge repository Next: Public-sector AI readiness indicators →