Malaysia - National AI Roadmap

Overview

The Malaysia National Artificial Intelligence Roadmap 2021-2025 (AI-RMAP) is a government strategy playbook prepared under the Ministry of Science, Technology & Innovation (MOSTI) that sets the national trajectory for AI deployment, capability building and governance. The document frames AI as an economic and social enabler and organises actions around six strategies and 24+ initiatives designed to catalyse adoption across priority sectors. The roadmap is published as a living document and public resource (see the national playbook and publication pages such as AI-RMAP full PDF and the MOSTI announcement at MOSTI news) and emphasises collaboration across government, industry, academia and civil society to achieve measurable adoption, talent and infrastructure targets by 2025. The roadmap highlights 11 national use cases across five priority sectors and embeds seven responsible-AI principles that later informed Malaysia’s National Guidelines on AI Governance & Ethics.

Definitions

AI-RMAP adopts an operational definition of artificial intelligence as a suite of technologies that enable machines to demonstrate intelligence, adapt to new circumstances and amplify human ingenuity. The roadmap distinguishes between components such as sensing technologies, advanced intelligent systems, analytics, machine learning, natural language processing and robotics. It uses standard governance terms: AI-CIU (AI Coordination & Implementation Unit), AI Catalyst (innovation hub), AI SCORE (SME adoption rating), sandboxes/testbeds, and ‘‘living document’’ to convey iterative updates and stakeholder feedback loops. The roadmap also defines risk tiers implicitly by sectoral impact, identifying higher-sensitivity domains (healthcare, public services, critical infrastructure) where additional controls, testing and oversight are recommended.

Governance and Institutional Framework

AI-RMAP recommends a centralised yet collaborative governance architecture anchored by MOSTI and an AI Coordination & Implementation Unit (AI-CIU) to oversee delivery, prioritisation and alignment of AI initiatives across ministries and states. The governance model is explicitly quadruple-helix, calling for co-design and co-funding across government, academia, industry and civil society, and it proposes specialised working groups for standards, ethics, procurement, data governance and sectoral deployment. The roadmap calls for a National AI Innovation Hub (AI-Catalyst) to host sandboxes, testing facilities and commercialization pathways and encourages the establishment of national datasets, compute infrastructure and linkages to the private sector (e.g., cloud providers). For ongoing oversight, AI-RMAP envisions regular national panels, stakeholder workshops and an online portal to collect feedback and publish progress; later governance outputs (for example, the National Guidelines on AI Governance & Ethics) were published by MOSTI to operationalise several of the roadmap’s governance and ethical recommendations (see MOSTI and MASTIC publications).

Key Focus Areas

The roadmap organises activity under six strategic thrusts. First, establishing AI governance, including ethical principles, national coordination (AI-CIU), standards development and policy roadmaps. Second, advancing AI R&D by strengthening academic-industry collaboration, funding fundamental and applied research, and institutionalising AI across national research institutes. Third, escalating digital infrastructure to enable AI by investing in data platforms, shared compute, national datasets and connectivity. Fourth, fostering AI talent through education, reskilling, certification programmes and public-sector training initiatives (targeting civil servants, SMEs and industry). Fifth, acculturating AI by raising awareness, supporting SME adoption via the AI SCORE program and public engagement. Sixth, kick-starting an AI innovation ecosystem via AI parks, innovation hubs, incentives and procurement pathways. The roadmap identifies 11 national use cases with technology enablers (sensing, analytics, robotics) and articulates sectoral goals for agriculture & forestry, healthcare, smart cities & transport, education and public services, while also encouraging cross-cutting investments in standards, ethics and testing facilities to reduce adoption friction.

Implementation Framework

AI-RMAP provides a phased implementation framework (Horizon 1–3) with specific initiatives linked to responsible-AI principles. Implementation levers include public funding for R&D and testbeds, incentives for private investment, partnerships with cloud and enterprise providers, SME support programs (AI SCORE), talent pipelines (training, micro-credentials, academic curricula) and procurement reforms to accelerate government uptake. The roadmap stresses the use of sandboxes and testbeds for high-risk and sectoral applications (e.g., healthcare diagnostics) and recommends standards and conformity processes to enable safety, interoperability and market confidence. It proposes an online portal for continuous stakeholder input and a playbook approach to replicate successful local pilots nationally. Key institutional links include MOSTI, MDEC, MIMOS and state-level actors working under coordinated committees such as the National Blockchain and Artificial Intelligence Committee (NBAIC).

Monitoring and Evaluation

Monitoring is envisaged through measurable KPIs tied to adoption, talent growth and infrastructure milestones. AI-RMAP sets targets for the number of SMEs adopting AI, the establishment of AI parks and the development of national datasets while requiring annual progress reports, stakeholder panels and updates to the living document. The roadmap calls for independent evaluation of high-impact pilots, public reporting of results and maintenance of a central monitoring dashboard to show KPI progress. Later public reporting on implementation (for example parliamentary updates and media reporting) indicates periodic progress checks and public accountability mechanisms tied to the roadmap's timeline.

Penalties, Liability, and Appeals

AI-RMAP itself is a non-binding strategic document and does not prescribe penalties, statutory liability regimes or administrative sanctions. Instead, it recommends that Malaysia develop sectoral regulatory instruments and standards, aligned to the roadmap and ethical principles, where necessary. For high-risk applications (healthcare, critical infrastructure, public services) the roadmap recommends conformity testing, certification and procurement safeguards; legal liability and redress are to be determined by future sectoral laws, standards and regulatory frameworks. Subsequent policy developments (for example, amendments to data protection law and cybersecurity legislation) have started to build the enabling legal architecture for enforcement and liability related to technologies recommended by the roadmap.

Relationship to Other Instruments

AI-RMAP is designed to align with existing national blueprints such as the National Digital Economy Blueprint (MyDIGITAL), the National Policy on Science, Technology & Innovation (DSTIN 2021–2030) and sectoral strategies (education, healthcare). It also connects to provider-driven ecosystems (MDEC-led initiatives) and standards bodies such as the Standards Malaysia mirror committees coordinated by national research agencies. The roadmap is intended to feed into and be operationalised by subsequent instruments including governance guidelines (AIGE), standards, sectoral regulatory updates and potential future AI-specific legislation.

International Alignment

The roadmap calls for alignment with international norms and references global instruments and best-practice frameworks (UNESCO, OECD, ISO) when designing ethics, standards and governance measures. It expressly recommends benchmarking against international codes and invites participation in global fora. Malaysia’s later publication of the National Guidelines on AI Governance & Ethics explicitly drew on UNESCO and OECD principles, demonstrating the roadmap’s intent to align domestic policy with international guidance and facilitate cross-border AI collaboration and trade.

Implementation Timeline

Sources and References