In March 2026, Vietnam began implementing its state-led, top-down AI law (Law 134/2025/QH15), one of the boldest moves to govern the juggernaut technology sector that any Southeast Asian government has made so far. For a country rapidly ramping up its high-tech ambitions, Vietnam’s AI regulatory measure aims to strengthen sovereign control over data flows, augment transparency, and mitigate against varying risks associated with AI. Moreover, it sets a benchmark for regulation within a region which has generally relied on voluntary compliance, as exemplified by ASEAN’s soft-law or guideline approach to AI governance.

But therein lies the rub. While Hanoi is forcefully implementing a standalone legal framework to assert control over information technology and strengthen state sovereignty, the Vietnamese government is ignoring the broader regulatory, socio-economic, and environmental implications. These include unintended consequences such as the sorts of economic dislocations and social disruptions that are inherent in such dynamic and historic transformations, although AI engenders change at exponential rates that could vastly widen already troubling societal inequalities.

Vietnam is not alone in its quest for enhanced digital sovereignty, and the stakes of the issue are high. The United States, China, Russia, and the European Union have long understood that key sectors – ranging from healthcare to energy and security – now operate on AI systems generally dominated by a handful of tech companies and the supply chains of critical minerals on which they rely. AI governance, therefore, increasingly encompasses the grand strategies of various global, national, and corporate stakeholders who are seeking to survive in a harshly competitive environment.

Vietnam’s new AI law creates a comprehensive legal framework, establishing a three-tier system of high, medium, and low categories of risk that align with differing levels of regulatory control. The higher the risk, the more government regulation. As a state-led approach, the law applies to both domestic and foreign entities conducting AI activities in Vietnam. While the law is still being implemented, it ideally aims to balance regulation with innovation and socio-economic development by allowing tech developers to test evolving AI models without strict penalties during a trial phase.

Although the law has its strengths, Vietnam’s AI law has several important shortcomings. First, the policy sits uneasily under ASEAN’s regional voluntary approach to AI regulation. Most troubling, it possibly creates a high-tech race to the bottom in the region, as venture capital often seeks jurisdictions with the least amount of regulation. Vietnam’s growing economy and quest for capital sits in a region increasingly shaped by ASEAN’s evolving governance architecture.

But the country’s AI law currently stands in contrast to ASEAN’s soft approach to AI regulation, at least for now. In February 2024, ASEAN issued its Guide on AI Governance and Ethics, which outlined principles of transparency, fairness, accountability, human-centricity, privacy, safety, and reliability. This framework was expanded in January 2025 to cover generative AI with nine areas of concern, including cybersecurity and content creation. ASEAN’s principle-based AI approach tends to mirror the EU’s emphasis on constructive engagement with the AI sector. By the end of 2026, ASEAN could potentially be moving to a legally binding approach under the ASEAN Digital Economy Framework Agreement, which is expected to be signed later this year.

Vietnam’s AI law is also notable for what is omitted from regulation, such as the questions of how to manage environmental degradation and natural resource allocation. This omission is significant because AI requires vast resources, particularly water for cooling data centers. According to the World Economic Forum, the standard 1 megawatt (MW) data center can guzzle approximately 25.5 million liters of water per year; this is comparable to the daily consumption of 300,000 residents in a medium-sized city in Vietnam. The energy implications of such resource consumption are stark. Diverting natural resources to cool AI’s growing infrastructure threatens to deplete local and regional water reserves. This strain on Vietnam’s environment is escalating in what the World Bank already views as one of the most at-risk countries for global climate change and rising coastal waters.

More importantly, the new law fails to offer any social protections for workers and communities who are displaced by AI technology. Across Asia, for example, workers in the gig economy have protested companies that control AI-generated algorithms, which have tightened delivery times while cutting wages. Vietnam has clearly missed an opportunity to take a more worker-friendly, human-centric approach to AI regulation. It is not alone in this; governments across the region have generally shirked any meaningful effort to provide a social safety net to support workers from a growing wave of AI systems and robotic automation.

These shortcomings collectively underscore how during periods of great transformation, various stakeholders, including countries, companies, and local communities, must cooperate to protect two essential commodities that are central to the global economy; namely, land and labor. Consequently, Vietnam’s AI law, and others like it that fail to address governance gaps involving key commodities and regulatory coordination, will undermine AI’s overall effectiveness, exacerbating economic dislocation and social displacement that could potentially unleash regional instability and uncertainty for many years to come.