Vietnam Launches National Portal for AI Risk Classification

Vietnam now has a working, centralized gateway for enforcing its artificial intelligence law. On August 28, 2026, the Ministry of Science and Technology launched the portal ai.dda.gov.vn, through which companies, AI system developers and government agencies are required to register their products for risk classification before they reach the market.

The portal functions as a single administrative window. Through the same platform, companies apply for controlled testing of their systems, submit risk classification results and register products in the national AI database, report serious incidents, and file periodic reports on how their systems are operating.

How the classification works

Vietnam's law takes a risk-based approach similar to the EU AI Act. High-risk systems are those that could seriously threaten life, health, citizens' rights or national security. Medium risk covers situations where a user could be misled by an undisclosed interaction with AI or by AI-generated content. Everything else falls into the low-risk category.

A separate unacceptable-risk category covers systems deemed a threat to national security, human dignity or public order. Providers must submit their own risk classification through the portal before deploying a system, under a self-declaration model. Once submitted, the portal automatically generates an AI system identification code and an electronic registration confirmation.

Differences from the AI Act

Tech regulation lawyers describe Vietnam's law outright as an example of the so-called Brussels effect, where other countries copy the structure of EU rules. The risk classification model and the conformity assessment requirements for high-risk systems are largely modeled on the AI Act.

There are notable differences, though. The EU AI Act lists high-risk system categories exhaustively, in annexes to the regulation. Vietnam took a different route, delegating the drafting of that list to the prime minister, which gives the government more flexibility to add new categories on an ongoing basis without amending the law itself. Penalty caps for violations are also lower in Vietnam than in the European Union.

Purpose of the portal

According to the ministry's statement, the portal is meant to ensure transparency and support the safe, responsible development of Vietnam's AI ecosystem. It publishes conformity assessment results and information on detected violations, giving the public a tool to monitor which systems are operating legally on the market.

The portal is meant to ensure transparency and promote the safe, responsible development of Vietnam's AI ecosystem - statement from Vietnam's Ministry of Science and Technology

The ministry also highlights a practical benefit for companies: moving all procedures into one electronic system is meant to cut the time and cost of meeting legal requirements. Instead of filing documents with several agencies, a business handles the paperwork in one place, and the portal also connects it to support programs, computing infrastructure and shared datasets available to AI developers.

Significance for the region

Vietnam is one of the first Southeast Asian countries with a comprehensive, standalone law regulating artificial intelligence. For readers in Poland, it is further evidence that the risk-tier approach to AI regulation, familiar from the EU AI Act, is becoming a global reference point, including outside Europe and the United States.

For technology companies operating internationally, it means another compliance regime to track, alongside the EU AI Act and rules being prepared in other Asian countries. Foreign firms offering AI systems in Vietnam will need to appoint a local representative and register their products through the new portal before entering the market there.