Every regional summit now has an artificial intelligence agenda, and many of those agendas consist of a declaration and a photograph. Southeast Asia has a more specific problem than most: the region's economies are deeply integrated into global technology supply chains as assemblers, testers and increasingly as designers, while its governments have limited capacity to regulate what those chains produce.
Why the region is exposed in a particular way
Three characteristics shape the exposure.
The first is the outsourcing economy. Business process services, content moderation, data annotation, medical transcription and software support are substantial employers in the Philippines, Malaysia and Vietnam. These tasks are directly in the path of automation: not because the work is trivial, but because it is structured, documented and repetitive enough to be partially automated.
The second is the semiconductor and electronics supply chain. Malaysia, Singapore, Vietnam, Thailand and the Philippines all host assembly, testing and design activity. AI accelerates demand for certain chips and specialises the supply chain, which rewards countries with skills and reliable infrastructure.
The third is language. English proficiency gave the region an advantage in the first wave of digital services. In an AI-mediated market, that advantage narrows, because the technology crosses language barriers more cheaply than people do.
Governance is being assembled quietly
ASEAN published a Guide on AI Governance and Ethics in 2024, aimed at voluntary adoption and written for a region with very different regulatory capacities. It is guidance rather than law, and it is deliberately light because a binding regional AI statute would be vetoed by several members.
National approaches diverge, which is the practical reality that regional frameworks must accommodate: Singapore has advanced an assurance and testing approach and issued detailed governance frameworks; Vietnam has issued rules on AI systems and data; the Philippines has been drafting legislation through its competition and trade agencies; Indonesia and Thailand have taken their own routes.
The likely regional outcome is interoperability rather than uniformity: common principles, mutual recognition of testing regimes, and shared red lines on a small number of high-risk uses.
The workforce question, honestly
The claim that a region of 650 million people will simply be displaced is too strong, and the claim that nothing will change is too comfortable. What is defensible is that the task composition of many service jobs will shift toward judgment, exception handling and client relationships, and that the premium on those skills will rise while the premium on throughput falls.
For the Philippines, this intersects with an outsourcing sector that employs well over a million people and contributes substantially to national income. The adaptation is not primarily a technology problem. It is a training, supervision and pricing problem: whether firms can move up the value chain faster than the automation arrives.
Compute, data and the missing infrastructure
AI capability depends on compute, data, power and connectivity. Southeast Asia is short of all four relative to its ambition. Data-centre investment is arriving in Malaysia, Indonesia, Thailand and the Philippines, driven by demand and by incentives, but it is constrained by electricity supply, water for cooling, and grid reliability.
This makes AI policy partly an energy policy. A country that cannot guarantee reliable power cannot host the infrastructure on which its own digital economy will depend, whatever its regulatory framework says.
The strategic dimension
AI is also becoming a foreign policy file. Export controls on advanced chips, standards around model deployment, and competition over undersea cables and cloud infrastructure all shape what smaller states can access. ASEAN's interest is in avoiding a bifurcated technology order in which members must choose a stack, because choosing is precisely what the bloc's method is designed to avoid.
That is the substance behind the summit language about a rules-based digital environment. It is not an abstraction: it is about whether a Philippine hospital, bank or university will be able to buy the tools it needs from more than one supplier.
A realistic regional agenda
What ASEAN can plausibly do in the next few years is unglamorous and useful: shared testing and evaluation capacity, common procurement standards, mutual recognition of AI assurance, regional datasets for underrepresented languages, and coordinated training programs. What it cannot do is regulate frontier development, which happens outside the region.
The summit will produce language on AI. The test is whether any of it converts into capacity that a member state can use: a testing lab, a shared standard, a recognised certification. Without that, the region will continue to consume rules written elsewhere.







