The roadmap observed that Malaysia lacked a central AI governance coordination structure, with most AI activity planned and implemented in silos. Strategy 1 argued that this siloed approach undermines competitiveness and public-service efficiency, and called for a clear national AI policy and a whole-of-nation approach rather than a compartmentalised one. AI governance was to be aligned with the Shared Prosperity Vision 2030 and the broader Sustainable Development agenda.
The AI Coordination and Implementation Unit (AI-CIU)
Central to Strategy 1 was the proposed AI-CIU — envisioned as the apex government body on all matters related to AI, answerable to the Minister for Science, Technology and Innovation, and designed as a lean, adaptive and responsive organisation. Its first task would be to establish a Foresight Committee for horizon scanning and policy advocacy, drawing on MIGHT, MOSTI's Technology Foresight division and outside experts.
Committee structure
The roadmap set a Horizon 1 (2021–2022) target of staffing the AI-CIU and standing up six committees — Policy & Regulation; Ethics; Talent; R&D&I; Data Sharing; and Communication & Social Media — together with 20 expert groups, with MOSTI as lead agency alongside MITI, KKMM and MDEC.
Strategic initiatives under Strategy 1
- 1.1 — Establishing the AI-CIU to drive implementation of the roadmap.
- 1.2 — Establishing a digital platform for multidirectional committee interaction and horizon scanning.
- 1.3 — Institutionalising cybersecurity policies for AI implementation, targeting 100% of ministries and agencies adopting AI cybersecurity policies.
- 1.4 — Institutionalising AI principles, including an AI Code of Ethics and guidelines disseminated to all stakeholders.
- 1.5 — Establishing clear guidelines for data sharing across government to enable AI implementation.
Principles for Responsible AI
The roadmap set out seven principles intended to govern AI design and deployment across sectors:
- Fairness — consistent recommendations for people in similar circumstances.
- Reliability, safety and control — dependable operation even under unexpected conditions.
- Privacy and security — compliance with privacy law and protection of personal data.
- Inclusiveness — addressing a broad range of human needs and capabilities.
- Transparency — openness about how data is used.
- Accountability — those who design and deploy AI remain answerable for it.
- Pursuit of human benefit and happiness — AI as a tool for human well-being.