National AI Office

A successful first Global Dialogue on AI Governance, from Malaysia's perspective, would be defined by practical, implementable outcomes rather than broad consensus statements. First, it should produce clear pathways from principles to practice, including usable approaches to risk classification, audit, and assurance that can be integrated into existing sectoral regulation. Second, it must recognise different national starting points, ensuring governance models are proportionate and do not impose undue burdens on emerging ecosystems like Malaysia's while still enabling alignment with global standards. Third, success would include tangible mechanisms for ongoing cooperation so that participation translates into sustained capability, not just dialogue. Ultimately, success lies in establishing AI governance as an iterative, cooperative process that balances risk management with innovation, while strengthening Malaysia's ability to participate competitively and responsibly in the global AI ecosystem.

• Safe, secure and trustworthy AI
• AI capacity-building
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Transparency, accountability, and human oversight

Safe, Secure, and Trustworthy AI: Trust is the bedrock of Malaysia's digital economy. Following the launch of the National Guidelines on AI Governance and Ethics (AIGE), our priority is to operationalize these standards. We aim to ensure that AI systems deployed within public institutions and the private sector are resilient against risks while protecting data sovereignty and citizen privacy. AI Capacity-Building: To compete globally, Malaysia is aggressively building local capabilities. Our mandate includes training 200,000 AI-skilled workers by 2030. By prioritizing capacity-building, we move beyond mere adoption toward high-value innovation, ensuring our workforce is equipped to lead in AI research, development, and commercialization within the regional ecosystem. Social, Economic, Ethical, Cultural, Linguistic, and Technical Implications: In line with Malaysia's multicultural identity, we prioritize AI that is inclusive and representative. This includes addressing algorithmic bias and ensuring that AI development respects our cultural and linguistic diversity. We are committed to a "human-centric" approach where technological progress does not widen social inequities but instead drives sustainable, shared prosperity for all Malaysians. Transparency, Accountability, and Human Oversight: We believe AI must remain human-centric. Malaysia is committed to establishing clear frameworks where AI decisions are explainable and accountable. By ensuring "human-in-the-loop" oversight, we mitigate risks of autonomous errors and ensure that AI remains a tool for empowerment rather than a replacement for human judgment and responsibility

The current AI governance landscape is transitioning from voluntary frameworks toward structured, enforceable legislation to address emerging risks such as deepfakes, algorithmic bias, and high energy consumption. While this evolution presents opportunities for Malaysia to leverage its large working-age population and lead regional AI safety initiatives, significant challenges remain regarding regulatory fragmentation and the technical opaqueness of AI systems. The primary challenge is regulatory fragmentation, as current sectoral guidelines lack a unified baseline and binding force. Existing non-binding frameworks like the AI Governance and Ethics (AIGE) guidelines are insufficient to manage escalating risks, including AI-related harms like forged documents that bypass security systems. Furthermore, technical risks like "black box" algorithms and algorithmic bias lead to discriminatory outcomes in finance and recruitment, while advanced models place immense pressure on national energy and water resources. Conversely, Malaysia is uniquely positioned to turn these gaps into opportunities through demographic advantages and strategic regional leadership. During the 2025 ASEAN chairmanship, the country is spearheading the "ASEAN AI Safe" network to harmonize regional standards. Domestically, the development of the AI Governance Act (targeted for end 2026, parliament presentation) provides a chance to establish a governance framework within the country.

The AI Dialogue can advance international cooperation by moving beyond principles toward shared, practical approaches to governance, especially in areas like risk assessment, testing, and transparency. This could help reduce fragmentation and gives clearer signals across jurisdictions. It also enables capacity-sharing, allowing countries to access expertise, tools, and regulatory practices needed to implement AI governance effectively. More importantly, it could establish ongoing cooperation channels such as knowledge-sharing so collaboration continues beyond the Dialogue itself. Overall this can help countries like Malaysia can adopt relevant standards while supporting domestic innovation.

From Malaysia's perspective, the AI Dialogue can advance international cooperation by reinforcing a principles-based yet practical approach to governance. It can translate shared values into common implementation practices, enabling interoperability without requiring rigid harmonisation. The Dialogue is also critical for capacity-building and access to expertise, which Malaysia needs to operationalise its largely guidance-driven framework and strengthen its evolving AI governance ecosystem. At the same time, it should support flexible alignment, allowing Malaysia to adopt and embed global standards and benchmarks without constraining innovation or imposing disproportionate burdens. Ultimately, the Dialogue provides a platform for Malaysia and other Global South countries to shape governance norms, ensuring they reflect development, inclusion, and economic priorities—not just frontier market concerns.

From Malaysia's perspective, the AI Dialogue can advance international cooperation by reinforcing a principles-based yet practical approach to governance. It can translate shared values into common implementation practices, enabling interoperability without requiring rigid harmonisation. The Dialogue is also critical for capacity-building and access to expertise, which Malaysia needs to operationalise its largely guidance-driven framework and strengthen its evolving AI governance ecosystem. At the same time, it should support flexible alignment, allowing Malaysia to adopt and embed global standards and benchmarks without constraining innovation or imposing disproportionate burdens. Ultimately, the Dialogue provides a platform for Malaysia and other Global South countries to shape governance norms, ensuring they reflect development, inclusion, and economic priorities—not just frontier market concerns.

These include women, children, and persons with disabilities (OKU communities), whose experiences with AI systems often differ significantly from dominant design assumptions. For example, bias in datasets and models can reinforce structural inequalities, while accessibility gaps in AI-enabled services can exclude OKU communities from fully participating in digital economies. Children are also uniquely exposed, as AI increasingly shapes education, information ecosystems, and online safety environments without always being designed with age-appropriate safeguards. Beyond these groups, informal workers, rural populations, and small enterprises are also often underrepresented.

Generally, the Dialogue could benefit from engagement formats that move beyond traditional panel discussions toward hands-on, problem-solving, and scenario-based interactions that reflect how AI governance functions in practice.

To promote effective AI governance and address its challenges, Malaysia has developed a multi-layered approach that connects global standards, national capacity-building, and practical implementation mechanisms. A key element is the development of the National AI Standards portal (myAI Standards) under the National AI Office, which consolidates more than 80 international standards into a single reference system. This helps translate fragmented global standards into an accessible national repository for regulators and industry. Malaysia also treats international standards as a form of "soft law" implementation layer, particularly through embedding international standards such as ISO/IEC 42001 (AI Management Systems) and ISO/IEC 23894 (AI Risk Management) within current and developing AI Policy. These standards operationalise ethical principles into auditable governance and risk processes that can be applied across sectors. In addition, Malaysia is part of the ISO Global Capacity-Building Journey, which focuses on strengthening regulatory capability through standards-based governance. This supports regulators in embedding technical and compliance expertise directly into policy implementation. These efforts are reinforced by the National AI Action Plan 2030 (AI Nation 2030), which provides a structured national roadmap to transition Malaysia into a value-adding AI producer. It includes: Impact engines across sectors such as healthcare, manufacturing, public services, and MSMEs Foundational enablers covering governance, infrastructure, talent, innovation, and funding Within this, the Responsible Governance pillar is particularly relevant, embedding an "AI Act and Trust Function" to ensure safety, accountability, and trust are integrated into system design and deployment. Together, these initiatives reflect a governance model that is standards-driven, capacity-focused, and implementation-oriented, ensuring AI governance is not only principle-based but operationally embedded across Malaysia's digital ecosystem.