Vietnam Enforces Southeast Asia's First AI Law

A Regulatory First for the Region

On 1 March 2026, Vietnam became the first country in Southeast Asia to enforce a standalone law governing artificial intelligence. The Law on Artificial Intelligence (No. 134/2025/QH15), passed by the National Assembly in December 2025 with 90.7% approval, introduces a risk-based framework covering the research, development, deployment, and use of AI systems across the country.

The timing matters. While neighbours like Singapore, Malaysia, and Thailand have relied on voluntary guidelines and soft governance frameworks, Vietnam has moved straight to binding legislation. That puts it alongside South Korea, which enforced its own AI law in January 2026, as one of only two countries in the Asia-Pacific with comprehensive AI regulation now in force.

What the Law Actually Requires

At its core, the law introduces a three-tier risk classification for AI systems: high, medium, and low. High-risk systems, those with potential to cause significant harm to life, health, or national security, face the strictest requirements. Providers must establish risk management mechanisms, maintain technical documentation, ensure human oversight, and register via a national AI portal before deployment.

The Ministry of Science and Technology (MoST) is preparing an official list of high-risk AI categories, expected later this year. Seven task groups have been launched to handle enforcement, training, and audits from 2026 onward.

Penalties are not symbolic. Serious violations can attract fines of up to 2% of a company's preceding year revenue. The maximum administrative fine sits at VND 2 billion (over $75,800) for organisations and VND 1 billion for individuals.

Transition Periods Give Businesses Breathing Room

Vietnam has not demanded overnight compliance. Existing AI systems in healthcare, finance, and education have until 1 September 2027 to meet the new requirements. All other sectors get until 1 March 2027. This phased approach mirrors the European Union's AI Act, which Vietnam's law draws clear inspiration from.

"AI serves as a support tool, and ultimate responsibility in critical social matters must remain with humans." - Nguyen Manh Hung, Minister of Science and Technology, Vietnam

The law also carves out space for innovation. A national AI development fund, controlled sandbox testing environments, and a voucher scheme for startups are all part of the package. These mechanisms echo approaches used in Japan and Singapore, balancing regulatory rigour with incentives for domestic AI growth.

How Vietnam Compares to Its Neighbours

Country	AI Governance Approach	Status (March 2026)	Enforcement
Vietnam	Standalone AI law (risk-based)	In force since 1 March 2026	Binding, with fines
South Korea	Comprehensive AI law	In force since January 2026	Binding
Singapore	Voluntary AI governance framework	Active since 2019, updated regularly	Non-binding
Thailand	AI ethics guidelines	Draft stage	Non-binding
Malaysia	Voluntary AI principles	Active	Non-binding
Indonesia	AI governance roadmap	Under development	Non-binding

The ASEAN Guide on AI Ethics and Governance, released in February 2024 with a supplementary guide for generative AI in 2025, provides a regional framework. But it remains voluntary. Vietnam's decision to go further with enforceable law reflects a calculation that voluntary guidance alone will not be sufficient as AI adoption accelerates across the region.

"Vietnam's first standalone AI law positions the country among early adopters in the region, alongside the EU and South Korea, in establishing comprehensive AI governance." - Trang Dang, Contributing Writer, International Association of Privacy Professionals (IAPP)

What Comes Next

MoST is now working on several implementing documents, including:

• A Prime Ministerial decision defining the official high-risk AI list
• Technical guidance for AI impact assessments and transparency requirements
• A National AI Ethics Framework to complement the legal requirements
• Decrees on specific sanctions and enforcement procedures

The law requires transparency for AI-generated content and mandates that providers of high-risk systems maintain operational logs. For foreign companies operating AI services in Vietnam, the compliance clock is now ticking, with the 12-month general transition period already underway.

For businesses already operating in Vietnam's fast-growing digital economy, the practical question is not whether to comply but how quickly they can map their AI systems against the new risk tiers. The law's IAPP analysis suggests that companies with EU AI Act compliance programmes will find significant overlap, but the Vietnamese requirements carry their own specifics around the national portal and local registration.

The AIinASIA View: Vietnam has made a deliberate bet that regulatory clarity will attract, not repel, AI investment. While the rest of ASEAN watches from the sidelines with voluntary frameworks, Hanoi has decided that binding rules create a more predictable environment for companies building AI products. The real test comes in execution. If MoST can publish the high-risk AI list and enforcement guidance before the transition periods expire, Vietnam could become the default regulatory reference point for the entire region. If the implementing documents stall, the law risks becoming a paper exercise. The next twelve months will tell us which version of this story we get.

Does Vietnam's AI law apply to foreign companies?

Yes. Any organisation that develops, provides, or deploys AI systems within Vietnam must comply with the law, regardless of where the company is headquartered. Foreign AI service providers operating in Vietnam face the same risk classification and registration requirements as domestic firms.

How does Vietnam's approach differ from the EU AI Act?

Both use a risk-based classification system, but Vietnam's law is shorter (36 articles versus the EU's 113) and includes innovation incentives like a national AI fund and startup vouchers. Vietnam's penalties are also calibrated differently, with a 2% revenue cap for serious violations compared to the EU's tiered fines reaching up to 7% of global turnover.

What happens if a company misses the compliance deadline?

Companies operating high-risk AI systems without proper registration and assessment face administrative fines of up to VND 2 billion for organisations. For serious violations, fines can reach 2% of the preceding year's revenue. The law also provides for potential suspension of AI system operations.

Will other ASEAN countries follow Vietnam's lead?

Thailand and Indonesia both have AI governance frameworks in development, but neither has signalled a timeline for binding legislation. Singapore has explicitly chosen to maintain its voluntary approach. Vietnam's experience over the next two years will likely influence how quickly its neighbours move toward enforceable AI rules.

Vietnam has drawn a line that no other Southeast Asian government has been willing to draw. Whether that line becomes a model or a cautionary tale depends entirely on what happens next. Is binding AI law the right move for developing economies, or does it risk slowing down the very innovation it claims to protect? Drop your take in the comments below.

YOUR TAKE

We cover the story. You tell us what it means on the ground.

Related Articles