Every AI Policy Ever · open in the interactive register
In forceBinding · Law
Law on Artificial Intelligence
Luật Trí tuệ nhân tạo
Record VN-02 · Law No. 134/2025/QH15
- Issued by
- National Assembly of the Socialist Republic of Viet Nam (15th National Assembly, 10th session)
- Jurisdiction
- Vietnam
- Level · region
- National · South & Southeast Asia
- Published / adopted
- 2025-12-10
- In force
- 2026-03-01
- Language
- Vietnamese
- Official identifier
- Law No. 134/2025/QH15
Summary
Law of 35 articles on the research, development, provision, deployment and use of AI systems and their state management. It classifies AI systems as high, medium or low risk: providers self-classify and notify medium- and high-risk systems, and high-risk systems need a conformity assessment before use. It prohibits uses such as deception or manipulation causing serious harm and exploiting vulnerable groups; requires disclosure of AI interaction and machine-readable marking and labelling of AI-generated content; sets duties for serious incidents; and provides for a national AI ethics framework and controlled testing. AI used solely for defence, security or cipher purposes is excluded.
Language
English translation by Viet Nam Government News (Government portal) Open it. The Vietnamese text is authoritative.
Safety topics
risk-based classificationtransparency & labellingprohibited practicesaccountability & liabilitypublic-sector use
Status notes
Passed by the 15th National Assembly at its 10th session on 10 Dec 2025, with 429 of the 434 deputies present voting in favour, and published in the Official Gazette (Công báo) No. 40. High risk is defined as possible significant harm to life, health, the lawful rights and interests of organisations and individuals, national interests, the public interest or national security; medium risk as the possibility of confusing, influencing or manipulating users who do not recognise that they are interacting with an AI system or with content it generated. Article 4 requires human control and the ability to intervene in all decisions and behaviour of AI systems. Article 5 encourages controlled testing of technology, risk-proportionate management and voluntary compliance mechanisms, and prioritises investment in data and computing infrastructure, safe AI, skilled people and shared AI platforms of national strategic character. Article 27 requires agencies that use AI in state management or public services, and operate a high-risk system or one with significant impact on human rights, social equity or the public interest, to prepare an impact assessment report. Article 29 makes the deployer of a high-risk system that was managed and used lawfully liable for any damage, with recourse against the provider or developer where agreed. Article 33 repeals Article 3(9), 4(7), 12(6), point đ of 34(2) and Chapter IV of Law No. 71/2025/QH15; Article 34 sets entry into force on 1 March 2026; Article 35 gives systems already in operation 18 months (health, education, finance) or 12 months (others) to comply, during which the state management agency may suspend or stop a system at risk of serious harm. Implementing instruments: Decision 367/QĐ-TTg, Decree 142/2026/NĐ-CP, Circular 05/2026/TT-BKHCN and Decision 33/2026/QĐ-TTg (VN-03 to VN-06).
Local copy
Archived. Last attempt 2026-10-14 10:15:04 UTC
Official source (third-party copy, not published) archive/VN-02/2026-10-12/official.pdf
Retrieved 2026-10-12 04:43:36 UTC
SHA-256 53be2f9993e5060cc0ce723fc506d6535c9358978dbbeff11324c8d6236cae69
Secondary source (third-party copy, not published) archive/VN-02/2026-10-12/secondary.html
Retrieved 2026-10-12 04:43:42 UTC
SHA-256 62b3ec18a4fcb5e55196e09c088085345694eff49e76c82a1904c0a033902bac
Verification
AThe instrument's own official text, or an official register, was stored and confirms the title, date and status. checked 2026-10-14 · official-url-fetched
Stored on 12 Oct 2026: the signed Law from the Government portal file server (datafiles.chinhphu.vn; 20 scanned pages, read through Tesseract OCR in Vietnamese, which drops or swaps some diacritics and reads the number as '134/2025/QHI5') and the Official Gazette page (congbao.chinhphu.vn). Stored on 14 Oct 2026: the English translation published by Viet Nam Government News on 15 Jul 2026 ('Below is full translation ...', ending 'For reference only') and the National Assembly newspaper's report of the vote of 10 Dec 2025 (daibieunhandan.vn). The PDF confirms the number, title, all 35 articles, Article 34 (entry into force 1 March 2026) and the closing formula (adopted by the 15th National Assembly, 10th session, on 10 Dec 2025, signed by Trần Thanh Mẫn); the gazette page confirms the number, issue date, effective date and Gazette No. 40; the newspaper confirms the vote of 429 of 434 deputies present. The summary is written from the Vietnamese text; the earlier 'treats data, computing and safe AI infrastructure as strategic national assets' misread Article 5(4). The government translation renders thử nghiệm có kiểm soát as 'regulatory sandbox' (Article 21) and 'controlled experimentation'; the summary keeps the literal 'controlled testing'. The Government portal document page (vanban.chinhphu.vn, docid 216334) was read on 12 Oct 2026 but not stored, and the gazette date of 22 Jan 2026 given in the earlier notes is not shown on the stored gazette page.
Second check: Independent check by Claude Code (claude-sonnet-5-5), 2026-10-14: 11 field(s) supported.
History
- 2026-10-14related: [{"slug": "vietnam-national-ai-strategy-2030", "id": "VN-01"}, {"slug": "viet-na → [{"slug": "vietnam-national-ai-strategy-2030", "id": "VN-01"}, {"slug": "viet-na — two-way link to VN-08, VN-09 added in run 2026-10-14-discover-leads-2026-10-142026-10-14-discover-leads-2026-10-14 · Claude Code (claude-opus-5-5)
- 2026-10-14english_version: – → {"kind": "government translation", "url": "https://en.baochinhphu.vn/viet-nams-l — Viet Nam Government News' English translation of the full Law is now a stored source.2026-10-14-refresh-translation-fixes · Claude Code (claude-opus-5-5)
- 2026-10-14notes: High risk is defined as possible significant harm to life, health, rights, natio → Passed by the 15th National Assembly at its 10th session on 10 Dec 2025, with 42 — Risk definitions rendered closer to Art. 9(1) (adds national interests and 'influence'); Art. 4 rendered literally ('all decisions and behaviour'); Arts 5(4), 27 and 29 added; the remark on which source was used before the signed text was found moved to verification.note; the gazette date of 22 Jan 2026 dropped because the stored gazette page does not show it.2026-10-14-refresh-translation-fixes · Claude Code (claude-opus-5-5)
- 2026-10-14summary: Viet Nam's first dedicated AI law, passed on 10 December 2025 by 429 of 434 depu → Law of 35 articles on the research, development, provision, deployment and use o — Rewritten from the Vietnamese text. The claim that the Law 'treats data, computing and safe AI infrastructure as strategic national assets' misread Art. 5(4), which only makes it state policy to prioritise investment in data and computing infrastructure, safe AI, skilled people and shared AI platforms 'mang tính chiến lược quốc gia' (of national strategic character); that provision is now in notes. 'First dedicated AI law' and the vote count, which the stored Law text does not show, were taken out of the summary (the vote moves to notes). Added from the text: self-classification and notification (Art. 10), conformity assessment (Art. 13), prohibited acts (Art. 7(2)), transparency and labelling (Art. 11), serious incidents (Art. 12), ethics framework (Art. 26), controlled testing (Arts 5, 21).2026-10-14-refresh-translation-fixes · Claude Code (claude-opus-5-5)
- 2026-10-14country: Viet Nam → Vietnam — Country names follow pipeline/frame.json, which uses common English short names; titles and issuing bodies keep the issuer's own spelling.manual-2026-10-14 · Claude Code (claude-opus-5-5)
- 2026-10-14jurisdiction: Viet Nam → Vietnam — Country names follow pipeline/frame.json, which uses common English short names; titles and issuing bodies keep the issuer's own spelling.manual-2026-10-14 · Claude Code (claude-opus-5-5)
Citation
National Assembly of the Socialist Republic of Viet Nam (15th National Assembly, 10th session). Law on Artificial Intelligence [Law No. 134/2025/QH15]. 2025-12-10. https://datafiles.chinhphu.vn/cpp/files/vbpq/2026/01/luat134.signed.pdf. Status: In force; verified 2026-10-14 (level A). Local copy archive/VN-02/2026-10-12/official.pdf retrieved 2026-10-12 04:43:36 UTC, SHA-256 53be2f9993e5060cc0ce723fc506d6535c9358978dbbeff11324c8d6236cae69. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record VN-02.
Improve this record
Know of an amendment, a status change or an error? Report news or an error on GitHub; the form opens with this record filled in.