Vietnam Law on Artificial Intelligence
Standalone AI law enacted 10 December 2025 and effective 1 March 2026: risk tiers, transparency duties for high-risk and generative systems, a national AI database, and sandbox provisions.
| Jurisdiction | Vietnam |
|---|---|
| Category | AI Regulations |
| Status | Active |
| Effective date | |
| Latest development |
Analysis
Vietnam’s Law on Artificial Intelligence No. 134/2025/QH15 is a standalone, risk‑based AI framework enacted on 10 December 2025 and effective 1 March 2026, establishing tiered risk categories, transparency and labeling duties (especially for high‑risk and generative AI), a national AI infrastructure and sandbox regime, and implemented via Decree 142/2026/ND‑CP and an AI ethics circular.Law No. 134/2025/QH15 Official PDF of Law 134/2025/QH15 Government summary – Vietnam.vn Six laws effective 1 March 2026 – Vietnam.vn IAPP overview
Key Requirements
1. Scope of application and covered activities
- The AI Law applies to research, development, provision, deployment and use of AI systems in Vietnam, and to related rights and obligations of organizations and individuals, and state management of AI activities. Official PDF of Law 134/2025/QH15 English summary of Law 134/2025/QH15 Baker McKenzie overview
- According to expert commentary, the law governs both domestic and foreign entities engaged in AI activities in Vietnam, including cross‑border provision of AI systems that are used in Vietnam. Baker McKenzie overview Duane Morris legal update GVW legal analysis
2. Risk‑based classification of AI systems
- Vietnam’s AI Law explicitly classifies AI systems into three risk tiers, focusing on outputs and uses of AI rather than model training alone. Regulayer summary of Law 134/2025/QH15 IAPP overview Pertama Partners compliance guide
- Commentaries identify three tiers:
- Prohibited / unacceptable‑risk AI (e.g., systems that seriously threaten national security, social order, or fundamental rights).
- High‑risk AI systems, subject to strict obligations and documentation.
- Other / lower‑risk AI, subject to baseline transparency and safety requirements. Regulayer summary of Law 134/2025/QH15 Novel provisions article – LuatVietnam Baker McKenzie overview
3. Obligations for high‑risk AI systems
- High‑risk AI systems must have safety documentation prepared before deployment, subject to post‑market audit rather than pre‑approval licensing. Regulayer summary of Law 134/2025/QH15 Pertama Partners compliance guide GVW legal analysis
- Commentaries outline core high‑risk obligations, including:
- Risk assessment and mitigation measures before and during deployment.
- Technical robustness and safety controls.
- Human oversight mechanisms to prevent or correct harmful outcomes.
- Incident reporting to authorities for serious AI‑related events. Novel provisions article – LuatVietnam Baker McKenzie overview Kim & Chang analysis
4. Transparency and labeling duties (including generative AI)
- Vietnam requires clear labeling of AI‑generated content, including deepfakes, where content could be indistinguishable from real images or recordings. France24 report on Vietnam AI law IAPP overview Pertama Partners compliance guide
- Providers and deployers must inform users when they interact with AI systems instead of humans, especially in conversational agents and virtual assistants. France24 report on Vietnam AI law Baker McKenzie overview Duane Morris legal update
- Commentaries describe generative AI obligations such as:
- Content origin disclosure and watermarking/labeling of synthetic media.
- Policies to prevent harmful, deceptive or unlawful AI‑generated outputs.
- Dataset transparency requirements for certain high‑impact generative systems. IAPP overview Pertama Partners compliance guide GVW legal analysis
5. National AI infrastructure and data/registry requirements
- The law and related policy documents support the creation of national AI infrastructure, including an AI computing center and enhanced data resources for AI, especially large language models in Vietnamese. France24 report on Vietnam AI law Vietnam.vn article on AI Law 2025 Six laws effective 1 March 2026 – Vietnam.vn
- Commentaries describe the establishment of a national AI database / registry of AI systems and providers, including registration duties for certain risk categories. Novel provisions article – LuatVietnam Regulayer summary of Law 134/2025/QH15 Baker McKenzie overview
6. Sandbox and pilot mechanisms
- The AI Law introduces sandbox provisions allowing controlled experimentation with innovative AI systems under regulatory oversight. Novel provisions article – LuatVietnam Baker McKenzie overview Pertama Partners compliance guide
- Legal analyses highlight that sandboxes target emerging high‑impact use cases (e.g., healthcare, financial services, smart cities) with temporary exemptions or tailored rules while maintaining safety and rights protections. GVW legal analysis Duane Morris legal update Kim & Chang analysis
7. Penalties and enforcement
- The AI Law and related decrees provide administrative fines and other sanctions for non‑compliance, with penalties reported up to VND 2 billion for certain violations. Pertama Partners compliance guide AP Law Japan note on Decree 142 Viet An Law update on Decree 142
- Implementation analyses emphasize post‑market supervision (audits, inspections, incident investigations) as the core enforcement mode, rather than heavy ex‑ante licensing. Regulayer summary of Law 134/2025/QH15 Baker McKenzie overview GVW legal analysis
Compliance Challenges
1. Interpreting and operationalizing risk tiers
- Organizations face difficulties mapping diverse AI use cases to the three statutory risk tiers, especially in sectors like finance and health. Baker McKenzie overview Novel provisions article – LuatVietnam Regulayer summary
- Industry guides report that many providers lack internal AI risk taxonomy and governance structures, making classification and documentation challenging. Pertama Partners compliance guide GVW legal analysis Duane Morris legal update
2. Building transparency and labeling across product lines
- News and legal coverage note that AI‑generated content labeling is technically and organizationally complex, especially for large platforms and multi‑modal generative tools. France24 report on Vietnam AI law IAPP overview Kim & Chang analysis
- Commentaries describe customer communication and UX challenges when informing users they are interacting with AI systems, especially in sectors where human contact is expected (e.g., banking support, healthcare triage). Baker McKenzie overview GVW legal analysis Duane Morris legal update
3. Data, infrastructure, and national database integration
- Reports highlight difficulties aligning enterprise data governance with national AI data initiatives and registries, including questions about data formats, API access, and privacy. Novel provisions article – LuatVietnam Vietnam.vn AI Law article France24 report
- Legal practice notes mention uncertainty around registration thresholds, reporting obligations, and interactions with existing sectoral registries. Baker McKenzie overview AP Law Japan note on Decree 142 Viet An Law update on Decree 142
4. Aligning AI ethics, governance and operations
- Circular‑level guidance introduces a National AI Ethics Framework, requiring alignment of corporate AI practices with principles such as fairness, accountability and transparency. [Kim & Chang analysis](https://www.kimchang.com/en/
Recent developments
- — Vietnam is preparing a draft tightening data security rules for international AI services, requiring AI providers to verify the legality of training data, digitally label AI-generated content, and curb misinformation affecting national security; the draft is slated for submission to the National Assembly at its October 2026 session[15]. (source)
- — A July 2026 legal update analyzes novel provisions of Law 134/2025/QH15 on Artificial Intelligence, highlighting Vietnam’s first comprehensive AI framework effective March 1, 2026 and suggesting further improvements to balance risk mitigation with technological advancement[13]. (source)
- — A May 2026 business-focused guide explains compliance with Vietnam’s AI Law No. 134/2025/QH15, outlining obligations for AI providers and deployers, risk-based classification, and the practical steps companies must take ahead of the 2027 grace-period deadlines[4]. (source)
- — Vietnam issued Decree 142/2026/ND-CP on April 30, 2026 (effective May 1, 2026) to guide implementation of the AI Law, detailing requirements that create a new regulatory landscape for tech companies and startups operating AI systems in the country[9]. (source)
- — An April 2026 compliance guide describes Vietnam’s AI Law as human-centric and risk-based, specifying enforcement powers including suspension, recalls, and fines up to 2% of annual revenue, and clarifying 12‑month and 18‑month grace periods for different AI sectors[12]. (source)
- — A March 2026 policy analysis notes that Vietnam’s AI Law, effective March 1, 2026, requires AI firms to self-classify systems as high, medium, or low risk, notify authorities for medium/high-risk deployments, and label AI-generated images, video, and audio, reflecting both innovation goals and tight state oversight[3]. (source)
- — A March 2026 law firm briefing calls Vietnam’s AI Law the first comprehensive, risk-based AI statute in Southeast Asia, outlining transitional relief (18‑month grace for finance/health/education and 12‑month for other sectors) and key compliance priorities for businesses[2]. (source)
- — Legal news coverage from March 2026 reports that Vietnam’s sweeping AI Law has come into force, introducing tiered risk categories and outright bans on “unacceptable” AI that poses systemic threats to national security or human dignity, drawing comparisons with the EU AI Act[11]. (source)
- — A late February 2026 article highlights that Vietnam’s AI Law, effective March 1, mandates human supervision of generative AI, clear labeling of AI-generated content and deepfakes, user notification when interacting with AI, and creation of a national AI computing center to support domestic AI capacity[7]. (source)
- — A February 2026 in‑depth legal analysis reviews the foundations of Vietnam’s AI Law and its outlook, emphasizing the March 1, 2026 entry into force, 12‑ and 18‑month grace periods, forthcoming detailed government decree, and strategic implications for AI providers and deployers in regulated sectors[1]. (source)
Related regulations
- Internet Information Service Algorithmic Recommendation Management Provisions — China, Active
- National Artificial Intelligence Policy Bill — Chile, Active
- Artificial Intelligence Liability Directive — European Union, Active
- Federal Law on the Development and Use of Artificial Intelligence — Mexico, Active
- Presidential Regulation on the Framework for Artificial Intelligence Development — Indonesia, Active
- Law Establishing the Framework for the Development and Use of Artificial Intelligence — Peru, Active
- California Consumer Privacy Act AI Amendments — California, Active, effective 2020-01-01
- GPAI Framework for Responsible AI — AI Regulations, Active
Put it into practice
- Generate the policy: NIST AI RMF policy generator (generatepolicy.com)
- Buy the policy pack: NIST AI RMF Implementation Policy (cyberpolicy.shop)
- Build it yourself: Pillar 06 Companion — The 2026 AI Risk Register (ciso.diy)
Browse all regulations · Compliance deadlines · Latest updates