Malaysia pushed AI safety at ADGMIN 2025. Five deliverables mattered

Malaysia went to Bangkok in January 2025 with a regional digital-policy agenda that stretched well beyond one AI announcement. At the fifth ASEAN Digital Ministers' Meeting, Digital Minister Gobind Singh Deo called for consistent digital and cybersecurity standards, particularly as AI systems spread across borders and sectors.[1][2]

The meeting approved or advanced several concrete regional items. These included a declaration to establish the ASEAN AI Safety Network, a guide for sustainable data-centre development, work against online scams, progress on the ASEAN Regional CERT and preparation for the next ASEAN digital masterplan.[2][3]

For Malaysians, the important distinction is between a document being adopted and a system becoming operational. A declaration can assign direction and owners; it does not by itself create a functioning safety network, prevent a scam, train a worker or make two government systems interoperable.[1][3]

The first deliverable was AI governance and safety. ASEAN's meeting report said ministers approved the declaration establishing the ASEAN AI Safety Network as a 2025 priority economic deliverable. It also called for stronger coordination across AI policy, governance, law, regulation, standards and engagement with industry.[2][3]

The second was cybersecurity cooperation. Malaysia said it would lead a one-year project to develop an ASEAN Cybersecurity Cooperation Strategy for 2026-2030. The joint statement also recorded progress towards operationalising the ASEAN Regional CERT and welcomed regional recommendations on countering online scams.[1][3]

The third was cross-border digital interoperability. Gobind argued that more consistent standards could help digital solutions work across ASEAN. This sat alongside Malaysia's support for completing negotiations on the ASEAN Digital Economy Framework Agreement, or DEFA, and the wider objective of a more connected regional digital market.[1]

The fourth and fifth deliverables concerned capacity and infrastructure: specialised 5G workforce programmes, and the ASEAN Guide on Sustainable Data Centre Development. Both are enabling layers for AI. Models need people who can deploy networks and facilities that can supply computing capacity without treating energy, water and resilience as afterthoughts.[1][2]

During the related meetings, Gobind met ITU representative Sulyna Abdullah. His public account said the discussion covered AI standards, regulatory and policy support, and Malaysian participation in ITU working groups through the newly established National AI Office.[1][4]

That matters because AI standards often become the hidden operating rules behind procurement, testing, risk reporting and cross-border compatibility. If Malaysia helps shape those rules early, local regulators and technology providers can prepare for common terminology and evidence requirements instead of adapting after standards have already settled elsewhere.[1][4]

The public record did not set out a timetable, named standards or published work programme for the NAIO-ITU discussion. It should therefore be read as an agenda for cooperation, not evidence that a standards programme had already been delivered.[1][4]

The ASEAN AI Safety Network needs a visible operating model: participating bodies, a secretariat or coordinator, the issues it can examine, how member states share evidence and what guidance it will publish. Without that, the public cannot distinguish an active safety mechanism from a standing declaration.[2][3]

The cybersecurity strategy needs similar specificity. A useful delivery register would show the lead agency for each workstream, dates for consultation and adoption, regional exercises, information-sharing arrangements and measures such as response time, participation and remediation. Scam-loss trends may be important, but they cannot be attributed to one strategy without careful evaluation.[1][3]

For the 5G workforce and sustainable data-centre items, the proof is more practical: course places, completion and competency results; which organisations adopt the guide; and reported energy, water, resilience or emissions measures. Publishing those basics would let businesses and citizens see whether regional diplomacy changed operating practice.[1][2]