Kuala Lumpur will host ASEAN's AI safety secretariat. What must happen next

Kuala Lumpur was selected to host the secretariat of the ASEAN AI Safety Network, placing Malaysia at the administrative centre of a new regional mechanism for AI governance and safety. The announcement followed the sixth ASEAN Digital Ministers' Meeting in Hanoi, held from 12 to 16 January 2026.[1][2][3]

The network did not begin with the Hanoi announcement. ASEAN leaders had adopted its founding declaration at the 47th ASEAN Summit in Kuala Lumpur on 26 October 2025. That declaration tasked Malaysia, as lead proponent, with coordinating the institutional and operational arrangements for ministerial endorsement.[4][3]

Choosing a secretariat location is a meaningful institutional step, but it is not the same as launching a fully staffed body. The public documents set out objectives and relationships; they did not provide the operating detail needed to judge capacity, independence, delivery or impact.[1][4]

The Ministry of Digital described AI SAFE as a platform for capacity building, regulatory preparedness and safeguards. ASEAN's joint statement placed its operationalisation alongside the existing Working Group on AI Governance and other regional initiatives, rather than creating a completely separate policy universe.[1][3]

A secretariat normally supplies the continuity that a rotating series of ministerial meetings cannot. It can maintain a work plan, organise technical groups, prepare meeting papers, track decisions, coordinate partners and publish outputs. For AI safety, those outputs could include evaluation methods, incident-learning formats, regulatory toolkits and shared training.[4][3]

The founding declaration also makes clear that the network is voluntary and non-binding. That design may help member states participate at different levels of readiness, but it also means influence will come from useful evidence, trusted convening and practical adoption—not from the secretariat issuing enforceable regional rules.[4]

Hosting the secretariat gives Malaysia day-to-day proximity to the network's coordination, partners and institutional memory. It can make Kuala Lumpur a regular meeting point for ASEAN officials, researchers, civil society and technology companies working on AI risk and governance.[1][4]

That position should not be confused with control over the region's AI policy. The declaration places the network within ASEAN mechanisms and calls for participation from member states and multiple stakeholder groups. Malaysia hosts and leads work on establishment, while regional legitimacy still depends on participation and endorsement across ASEAN.[4][3]

For Malaysian universities, safety researchers and technology providers, the opportunity is to contribute tested methods and evidence. The risk is that access becomes concentrated among a small group of large organisations. Transparent calls, published selection criteria and open summaries would make the convening role more credible.[4]

The most useful next document would state the secretariat's legal and governance structure, accountable leadership, staffing, host arrangements, funding sources and conflict-of-interest rules. The 2025 declaration specifically asked ministers to explore sustainable funding models, showing that financing was recognised as an open implementation issue.[4]

A first-year work plan should identify a small number of deliverables with dates and owners. Examples could include a regional capacity programme, an AI evaluation exercise, a common incident-learning template or guidance for regulators. Each output should state who used it and what changed after adoption.[1][3]

Public reporting matters because an international secretariat can stay busy without becoming useful. Meeting counts and partner logos show activity. Stronger evidence would show member-state participation, completed tools, assessed capability, policy adoption, lessons from exercises and corrections made after failures.[4][3]