Vietnam's AI Law: What Businesses Need to Know
- 2 days ago
- 1 min read
Vietnam's Law on Artificial Intelligence (No. 134/2025/QH15) — the country's first comprehensive AI legislation — took effect on 1 March 2026. It adopts a risk-based, "manage for development" approach: supporting innovation while ensuring human oversight, transparency, and accountability, and giving businesses and regulators a common set of rules as AI adoption grows.
It is supplemented by Decree 142/2026/ND-CP and Decision 33/2026/QD-TTg, which sets out the official high-risk AI list (effective 15 August 2026). The framework is phased, allowing time to prepare.
*Key points:*
* *Classify* each AI system by risk level before deployment; keep dossiers for medium- and high-risk systems.
* *High-risk list:* 46 systems across six sectors — education, ethnic and religious affairs, healthcare, banking, judicial proceedings, and transportation.
* *Deadlines* for systems already in operation: 1 September 2027 (healthcare, education, finance); 1 March 2027 (other sectors).
* *Ongoing duties:* portal notification, transparency and labelling of AI content, human oversight, and conformity assessment for high-risk systems.
* *Incident reporting:* within 72 hours for urgent or uncontrolled events; within 5 working days for other serious incidents.
One point worth noting: compliance does not, by itself, extinguish civil liability — sound governance and clear contractual allocation of responsibility remain essential.
Full analysis below.
Comments