On March 1, 2026, Vietnam’s Law on Artificial Intelligence (Law No. 134/2025/QH15) came into effect — making it the first comprehensive AI law enacted in Southeast Asia（Source：Vietnam AI Law Passed by National Assembly – The Legal Wire、Vietnam Government Portal (original Vietnamese text)）.

Many assume this law is irrelevant to them since it’s a Vietnamese regulation — but that’s not necessarily true. Conversely, not every company is required to take action. This article directly addresses the question: “We use a Vietnamese team for offshore development, but our clients are Japanese companies — does this law apply to us?”

First, identify which case applies to you

Whether Vietnam’s AI Law applies depends on two factors: “who the system is built for” and “are you using AI or building AI”.

Case ①：Outsourcing development to a Vietnamese team for a Japan-facing system

This is the most common model — Japanese SIers or enterprises hire Vietnamese engineers to develop systems used by Japanese end users.

→ In most cases, this does not fall under the direct scope of this law.

The law applies to “domestic and foreign organizations and individuals conducting AI-related activities within Vietnam’s territory”（Source：Vietnam Government Portal, Article 2 (original Vietnamese text)）. If a Japanese company is simply contracting development work for Japanese users, no legal obligations arise under this law.

The same applies to AI tools used during the development process (such as GitHub Copilot or Cursor). These are developer productivity tools and do not constitute “providing” an AI service to Vietnamese users, so they fall outside the law’s scope.

Case ②：Providing or deploying an AI-powered system to Vietnamese users

This applies when a Japanese company integrates AI features into services targeting the Vietnamese market, or develops and deploys custom AI models (RAG systems, fine-tuning, etc.) for Vietnamese users.

→ This falls within the scope of the law. Compliance action is required.

Specific examples include:

• Integrating an AI chatbot into an app offered to Vietnamese users
• Developing and deploying custom AI models (RAG, fine-tuning, etc.) for the Vietnamese market
• Offering AI-driven services in healthcare, financial screening, or educational assessment in Vietnam

Two things to know even if Case ① applies

Even if your primary focus is developing systems for Japan, these two points are worth understanding.

Design with future Vietnam expansion in mind

If you’re considering expanding to the Vietnamese or Southeast Asian market down the line, building compliance considerations into your architecture now can avoid costly redesigns later. For AI-enabled systems in particular, retrofitting risk classification requirements after the fact tends to be significantly more expensive.

Verify that your Vietnamese partner team is aware of the law

How a Vietnamese offshore team responds to domestic regulations reflects their credibility as a partner. Even if the law doesn’t directly apply to your engagement, a partner with low regulatory awareness tends to have looser risk management practices overall.

Case ②：What actions are required?

If your system is provided or deployed to Vietnamese users, the following steps are required.

STEP 1：Classify your AI system by risk level

The law classifies AI systems into three tiers based on risk, with corresponding obligations for each（Source：Vietnam’s First Standalone AI Law – IAPP）.

High risk (medical diagnosis, financial screening, educational assessment, critical infrastructure, government decision support, etc.)

Applies to AI with direct impact on human life and rights. Key obligations include:

• Pre-deployment review application to the government (a process to confirm absence of risk before going live)
• Notification and registration with the Ministry of Science and Technology (Article 10)
• Implementation of human oversight and correction mechanisms (Article 14)
• Immediate reporting to the designated government authority when incidents occur
• Foreign companies must establish a commercial presence or authorized representative in Vietnam (Article 14, Clause 6)

Medium risk (conversational AI that could mislead users or influence their behavior)

Many chatbot services are likely to fall into this category. Companies must conduct their own risk classification, maintain records, and explicitly notify users that “this service is powered by AI”（Source：Vietnam AI Law (134/2025/QH15) Overview – CAST Vietnam）.

Low risk (most other AI systems)

Obligations are minimal. Having the ability to explain the system when questioned is generally sufficient.

STEP 2：Clearly define “who is the provider” in your contracts

The law classifies parties involved with AI by role（Source：Vietnam’s First Standalone AI Law – IAPP）.

In many arrangements, the Vietnamese engineering team acts as the “developer” while the Japanese company is the “provider” — placing heavier obligations on the Japanese side. If your current outsourcing contracts leave these roles undefined, now is the time to clarify.

STEP 3：Label AI-generated content with an “AI mark”

Audio, images, and video generated by AI must carry an electronic tag indicating their AI origin（Source：Vietnam – The First Law on Artificial Intelligence – Duane Morris）. For chatbots, users must also be able to clearly identify that they are interacting with AI.

Existing systems：Check the transition deadlines

AI systems already in operation before the law took effect have been granted a transition period（Source：Vietnam: Artificial Intelligence Law, Article 35 – Baker McKenzie）.

New systems must comply immediately. Existing systems should begin preparation well before the applicable deadline.

Risk of non-compliance

Violations can result in suspension orders, corrective action requirements, and financial penalties (Article 29). Legal analyses indicate that the maximum fine for legal entities is VND 2 billion (approximately JPY 12 million, subject to exchange rate fluctuation), though specific amounts will be defined in implementing decrees（Source：Vietnam AI Law 134/2025 Compliance Guide – Pertama Partners）.

Not just regulation：Incentives worth leveraging

Vietnam’s AI Law also includes mechanisms that work in favor of businesses（Source：Vietnam AI Law: Regulatory Milestone – Vietnam Briefing）.

• Regulatory Sandbox: Open to foreign companies — allows testing of new AI technologies in a controlled environment
• National AI Development Fund: Financial support for AI research and development
• Startup Voucher Program: Subsidizes a portion of AI development costs
• AI Model Capitalization: Allows AI models and algorithms to be recognized as company capital

For companies considering Vietnam as a base for AI development, these incentives are worth exploring alongside compliance planning.

Summary

Detailed regulations will continue to be refined. Partnering with an offshore development company that has a local legal entity in Vietnam enables you to maintain a development framework aligned with the latest regulatory developments.

We will propose the best approach for your project

As a leading Japanese IT solutions company with approximately 20 years of experience in Vietnam offshore development, we provide software and system development services.
Please feel free to contact us when considering offshore development.

Vietnam Offshore Development

Case Studies | Download Materials | Free Consultation | Contact via LINE