This page preserves a reference summary of the seven Principles for Responsible AI from the National AI Roadmap 2021–2025, published by MOSTI (ISBN 978-967-19025-5-4) and reproduced for research and citation with attribution. The full text is in the Playbook PDF; the official copy is at mastic.mosti.gov.my. Independent four-year analysis of how this strategy fared is in Roadmap vs Reality.
Threaded through the entire roadmap was a commitment that Malaysia’s AI would be responsible by design. Rather than leave ethics as an afterthought, the AI-Rmap set out seven principles intended to govern how AI is designed and deployed across every sector — a Malaysian articulation of the responsible-AI norms then emerging worldwide.
The principles were meant to be operationalised through Strategy 1’s governance machinery: an Ethics committee, an AI Code of Ethics, and guidelines disseminated to all stakeholders.
The seven principles
- Fairness. AI should make consistent recommendations for people in similar circumstances, and avoid encoding or amplifying unjust bias.
- Reliability, safety & control. AI systems should operate dependably, including under unexpected conditions, and remain under meaningful human control.
- Privacy & security. AI should comply with privacy law and protect personal data against misuse and compromise.
- Inclusiveness. AI should address a broad range of human needs and capabilities, and not exclude or disadvantage groups.
- Transparency. There should be openness about how data is used and how AI systems reach their outputs.
- Accountability. Those who design and deploy AI remain answerable for how it behaves and the outcomes it produces.
- Pursuit of human benefit & happiness. AI should ultimately be a tool for human well-being and the public good.
How they fit the global picture
The seven principles closely track the responsible-AI frameworks of the period — the OECD AI Principles, and the fairness / reliability / privacy / inclusiveness / transparency / accountability framing used by major standards bodies and technology firms. That alignment was deliberate: it positioned Malaysia to interoperate with international norms rather than invent its own in isolation.
These voluntary principles were the seed of what became, under the National AI Office, a National Guidelines on AI Governance & Ethics (AIGE) and — expected at Cabinet in 2026 — a risk-based AI governance framework with statutory backing. The 2021 principles endured in spirit even as the institution carrying them changed. Compare the principles then & now →