AI Regulation Hub

Vietnam

Vietnam has adopted a dedicated Law on Artificial Intelligence (Law No. 134/2025/QH15), issued 10 December 2025 and effective 1 March 2026. It uses a risk-based approach with high-, medium- and lower-risk categories, transparency duties (including deepfake labelling), supervision, and corrective measures.

Key provisions

Law No. 134/2025/QH15 on AI

In force

National AI statute creating a legal framework for AI development, provision, deployment and use. Risk-based; covers transparency, risk management, supervision and accountability.

Risk-tiered scope

In force

High-risk AI may significantly affect safety, health, lawful rights or public interests. Medium-risk AI may create user confusion, manipulation or harmful reliance (especially where AI output is hard to distinguish from human content). Lower-risk AI sits below those tiers.

Transparency & deepfake labelling

In force

Users must be informed when interacting with AI (chatbots, virtual assistants, automated customer service). AI-generated audio, images and video that simulate real persons, events or contexts must be marked clearly. Especially relevant to deepfakes, synthetic media, AI avatars and voice cloning.

Supervision, corrective measures, penalties

In force

Authorities may require suspension, withdrawal or other corrective measures where AI creates unacceptable or unlawful risk. Violations may lead to administrative penalties, criminal liability and compensation depending on conduct and harm.

National AI ethics framework

In force

Addresses responsible AI research, development and application, including human oversight and human intervention according to level of impact. Reviewed periodically or on major AI changes.

Detailed overview

Vietnam has adopted a dedicated Law on Artificial Intelligence, Law No. 134/2025/QH15. The law was issued on 10 December 2025 and is scheduled to take effect on 1 March 2026. Vietnam is therefore one of the jurisdictions with a specific national AI statute, rather than only AI policy guidance or data-protection rules.

Vietnam's AI law establishes a national legal framework for AI development, provision, deployment and use. It applies to AI-related activities in Vietnam and creates rules for transparency, risk management, supervision and accountability. The law is connected with Vietnam's earlier national AI strategy, which was approved by the Prime Minister in 2021 and set the country's direction for AI research, development and application to 2030.

Risk-based approach

Vietnam uses a risk-based approach. AI systems are divided into risk levels, including high-risk, medium-risk and lower-risk systems. A high-risk AI system is an AI system whose operation may significantly affect safety, health, lawful rights or public interests. A medium-risk AI system may create risks such as user confusion, manipulation or harmful reliance, especially where users cannot easily distinguish AI-generated output from real human-created content.

Transparency and deepfakes

Transparency is a key requirement. Users must be informed when they are interacting with AI, including chatbots, virtual assistants and automated customer-service systems. AI-generated audio, images and video that simulate real persons, real events or real contexts must be marked or labelled in a way that is easy to recognise. This is especially relevant for deepfakes, synthetic media, AI avatars, voice cloning and AI-generated video content.

Supervision and penalties

Vietnam also requires supervision and inspection for AI systems that create higher levels of risk. Authorities may require suspension, withdrawal or other corrective measures where an AI system creates unacceptable or unlawful risk. Violations may lead to administrative penalties, criminal liability and compensation obligations depending on the conduct and harm caused.

National AI ethics framework

Vietnam has also adopted a national AI ethics framework. The framework addresses responsible AI research, development and application, including human oversight and the ability for human intervention where needed according to the level of impact. The framework is intended to be reviewed periodically or when major changes occur in AI technology and use.

Ready to launch legally?

Book a 30-minute consultation. We'll map your licensing path and tell you exactly what's required, in plain language.