Vietnam's Law on Artificial Intelligence No. 134/2025/QH15 was adopted by the National Assembly on 10 December 2025 and entered into force on 1 March 2026. The Law applies to domestic and foreign organisations that research, develop, provide, deploy, or use AI systems in Vietnam or whose AI systems produce effects on Vietnamese persons. AI systems already in operation before 1 March 2026 benefit from a transitional period of 12 to 18 months, depending on the sector, during which they must satisfy the Law's obligations. Systems posing serious risks are excluded from the transitional period.
Article 9 of the Law classifies AI systems into three risk tiers: high, medium, and low. High-risk AI systems must be registered with the Ministry of Science and Technology, undergo conformity assessment, and be subject to ongoing monitoring by their providers and deployers. Article 14 requires providers to ensure transparency and explainability appropriate to the system's risk tier. Article 27 requires providers of generative AI systems to disclose training data sources and apply digital watermarks or equivalent technical markers to AI-generated content. AI activities exclusively serving national defence, national security, or cipher purposes are excluded from the Law's scope.
AI system providers who supply products into Vietnam without a local entity are not exempt: the Law applies wherever AI systems produce effects on Vietnamese persons. Deployers of high-risk AI systems in healthcare, financial services, critical infrastructure, and employment contexts face intensive compliance requirements, including conformity assessment and post-market monitoring. Generative AI service providers must implement content labelling from 1 March 2026, with the transitional period not available to systems newly placed on the market after that date. Foreign technology companies distributing AI-powered products or services to Vietnamese users must classify those products and map applicable obligations before supply.
Implementing regulations for several provisions of the Law were still being finalised as of 1 March 2026, leaving interpretive uncertainty for AI system categories whose risk classification depends on subsidiary criteria not yet published by the Ministry of Science and Technology. The Law spans 35 articles across 8 chapters and does not define all system categories within the statutory text; guidance from sector-specific regulators is expected to follow. The 18-month transitional deadline for high-risk systems in certain sectors falls in September 2027, and the 12-month deadline for other categories falls in March 2027.
Licentium advises on AI regulatory compliance in Southeast Asia and can draw on a partner network for Vietnam-specific legal counsel. Organisations assessing the risk classification of their AI systems, designing compliance programmes under Law No. 134/2025/QH15, or evaluating the transitional provisions' application to existing deployments are welcome to contact us. Work we undertake includes AI system risk classification, regulatory compliance gap analysis, generative AI provider obligations, AI governance documentation, and market-entry regulatory strategy for technology companies.