Malaysia AI Safety Institute

The Malaysia AI Safety Institute (MY-AISafe) is a Malaysian national body established in July 2026 to strengthen the country's technical capability in assessing and managing risks arising from artificial intelligence. It operates within AI Malaysia Berhad, the corporation that institutionalised the National AI Office (NAIO), and supports Malaysia's declared ambition of becoming an "AI Nation" by 2030.^[1]

History and Background

Malaysia's national AI governance architecture began with the National AI Office, established on 12 December 2024 to coordinate the country's AI policy agenda.^[1] On 28 July 2026, at a launch ceremony in Cyberjaya, the Prime Minister, Dato' Seri Anwar Ibrahim, officially announced the establishment of AI Malaysia Berhad — a permanent national AI entity under the Ministry of Digital that formalised and expanded the NAIO's functions — together with the National AI Action Plan 2026–2030 and the creation of the Malaysia AI Safety Institute.^[2][3]

The institute was introduced as part of the action plan's digital-trust pillar. Ministry of Digital officials described MY-AISafe as a strategic initiative to strengthen national technical capabilities by leveraging expertise from government, industry, academia and research institutions, so that Malaysian regulators and organisations can better evaluate, manage and mitigate AI risks. Its creation coincided with the drafting of Malaysia's AI Governance Bill, a risk-based legal instrument opened for public consultation and targeted for completion by the end of 2026.^[2][4]

Mandate and Structure

MY-AISafe's work is organised around three pillars. The first, AI RDCI (Research, Development, Commercialisation and Innovation), covers safety evaluations, testing and AI red teaming of AI systems. The second covers strategic collaborations for talent development, expertise sharing and technology transfer at national and international levels. The third covers governance, safety and regulatory support through the development of standards, risk management frameworks and compliance assurance.^[1]

Within AI Malaysia Berhad, the institute is expected to act as the technical authority that regulators and deploying organisations consult when assessing AI risk, complementing the policy coordination carried out by the parent corporation. The design deliberately mirrors the network of national AI safety institutes established in other jurisdictions — such as the United Kingdom, the United States and Singapore — which conduct technical evaluations of frontier models and coordinate internationally on evaluation methodology.^[4][5]

Applications and Impact

MY-AISafe is intended to give Malaysia a domestic reference point for questions such as whether an AI system deployed by a bank, a government agency or a hospital is safe, tested and compliant. Its functions support sector-specific AI adoption, consistent with the Ministry of Digital's stated shift from a whole-of-government to a whole-of-nation approach in which industry, academia and civil society participate in risk discussions.^[2] The institute also underpins Malaysia's goal of positioning itself as a regional hub for AI safety and governance in Southeast Asia, and its international collaboration mandate is expected to include joint evaluations and expertise exchanges with counterpart institutes.^[1][2]

>See Also

• National AI Office Malaysia
• AI Regulation in Malaysia
• Malaysia AI Governance Framework
• AI Safety
• PDPA and AI Compliance
• ASEAN AI Governance