by Linh Dang (Policy Intern at CELI)
By the end of 2025, the Vietnamese National Assembly will adopt a comprehensive legal framework on AI, the first of its kind in Vietnam’s legal history, into official law.
This is not the first time Vietnam has explicitly expressed its ambition to regulate and master AI technologies through legislative tools. Four years prior, the Vietnamese government introduced a national strategy for AI R&D until 2030, which outlines the country’s dream to become a future global hub for AI research and innovation in the next decades.[1] Since June 2025, Vietnam has integrated its AI regulation principles into a separate law code on digital technology industry, in which the term AI (trí tuệ nhân tạo in Vietnamese) is formally defined as “machine-based systems designed to operate with varying levels of autonomy and adaptability after implementation for achieving clear or implicit objectives.”[2]
However, the most notable moves to prepare for the law code occurred in late 2024, when the Vietnamese Politburo issued the Resolution No.57-NQ/TW[3] (termed as one of the “Four Pillars” driving the national governance strategy until 2045) on breakthroughs in the development of science, technology, innovation and national digital transformation. Within the framework of the Resolution, data governance and AI are framed as key strategies for Vietnam’s transformation to a prosperous and developed nation by the middle of the 21st century. Around the same time, the Vietnam Software and IT Services Association (VINASA) founded the national AI Ethics Committee to oversee the ethical development of AI in the country and promote a supportive legal framework for AI.[4]
According to Oxford Insights, Vietnam stood at the 51st position out of 188 countries worldwide in the 2024 AI Readiness Index.[5] The index evaluates international governments’ progress in integrating AI into the public sector in three pillars: government, technology sector, and data & infrastructure. Vietnam’s overall scores put the country ahead of both the global average and the lower middle-income average across most indicators. However, Vietnam is lagging behind many regional powerhouses in AI, such as China, Singapore, and South Korea. Thus, the incoming AI Law, which is available as a complete draft version for nationwide consultation, is expected to boost the country’s AI profile and quickly bridge the gap between Vietnam and large players in the international AI arena.
Figure: Vietnam’s 2024 AI Readiness Index per dimensions and comparison with lower middle income average and global average (Extracted from Oxford Insights)
Vietnam’s prospective AI regulatory framework: comparison with EU AI Act
Vietnam is striving to catch up quickly with the developed world in the field of AI and seeks to harmonize its regulations with those of the international standards. The structure of the proposed AI law is remarkably similar in approach to the European Union’s (EU) 2024 AI Act but differs in its strategic objectives, with a strong emphasis on the state’s role and technological sovereignty.[6] [7]
- The state is the leader of strategic AI development in Vietnam, whereas the EU largely concentrates on market harmonization and international confidence. The government should ensure that Vietnam’s AI governance philosophy encompasses four principles mentioned in Article 4: human-centeredness (nhân văn), security (an toàn), self-sufficiency (tự chủ), and comprehensive and sustainable development (phát triển bao trùm và bền vững). Furthermore, it is tasked with leading the planning, investment, and mastering of National AI Infrastructure as well as ensuring data sovereignty and reducing over-reliance on foreign technology.
- The Vietnamese government adopts a risk-based mechanism to AI governance that follows the EU’s system of four levels of AI risk (Unacceptable, High, Limited, and Minimal/No). However, slight alteration to the categorization (Unacceptable, High, Medium, Low) signifies Vietnam’s general openness and flexibility for domestic AI innovation. Only systems that clearly pose serious harm are subject to mandatory compliance, while those with lower risks are encouraged to apply “voluntary” regulation standards.
- Both frameworks forbid ethically dangerous usage of AI that poses severe risks to the public and basic human rights, with narrowly defined exceptions. Vietnam’s draft law introduces a specific set of additional prohibitions, such as the ban on development and use of AI deemed as reactionary and harmful to the State, as well as the production or distribution of AI-generated content, including deepfakes, that poses a significant risk to national security, public safety and social order.
- The EU AI Act imposes stringent requirements on conformity assessment for high-risk systems before they can enter the market. In contrast, the Vietnamese approach defaults to the post-market mechanism (hậu kiểm), emphasizing continuous self-assessment and ongoing supervision. Mandatory pre-market conformity and impact assessment (tiền kiểm) only apply to a portion of high-risk systems that pose the most obviously serious risks. In such cases, the AI Impact Assessment (AIA) procedure entails locating, evaluating, and suggesting mitigation measures for any possible effects on society, ethical values, and human rights.
- Both the EU AI Act and Vietnam’s proposed AI law require providers of general-purpose AI models (GPAI) with systemic risks to undertake additional responsibilities, including model assessment and ongoing risk mitigation. However, Vietnam’s strategy diverges from that of the EU when it comes to the provisions of using copyrighted material for AI training. The draft law explicitly mentions that as long as the model provider respects the owner’s right to opt out via machine-readable signals, duplicating or extracting legally accessible copyrighted works for the exclusive purpose of training AI models is typically not regarded as copyright infringement. This does not only indicate the national leadership’s ambition to facilitate the development of made-in-Vietnam AI models but also reflects the country’s weaker institutional norms of copyright protection compared to Western states.
- The Vietnamese government will play an active role in providing financial assistance to strategic AI R&D projects through the establishment of a non-budgetary National AI Development Fund. Additionally, the government encourages public procurement of domestic AI solutions and supports SMEs through AI Vouchers.
Conclusion
Explaining the rationale for Vietnam’s future AI governance policy, Minister of Science and Technology Nguyễn Mạnh Hùng recognizes the potential effect of artificial intelligence as “mightier than nuclear power” and stresses that the prospective AI law will institutionalize its role as an essential part of Vietnam’s intellectual infrastructure.[8] Thus, the framing of the draft AI law in Vietnam as not only a regulatory tool but also the foundation of a globally reputed AI industry is, in itself, significant and strategic.
Vietnam’s AI regulation trajectory largely follows the internationally recognized risk-based pattern. The government flexibly integrates good practices championed by the EU AI Act while putting unique emphasis on digital sovereignty, domestic AI autonomy and state-led development of core AI infrastructure, which is consistent with Vietnam’s long-term development visions. Vietnam is at the door of setting itself apart as an innovative AI governance model in the Indo-Pacific region, and the draft AI law has given observers an interesting look into how that future may unfold.
[1] thuvienphapluat.vn. “Quyết định 127/QĐ-TTg 2021 về Chiến lược nghiên cứu phát triển và ứng dụng trí tuệ nhân tạo đến 2030,” January 26, 2021. https://thuvienphapluat.vn/van-ban/Cong-nghe-thong-tin/Quyet-dinh-127-QD-TTg-2021-Chien-luoc-nghien-cuu-phat-trien-va-ung-dung-Tri-tue-nhan-tao-den-2030-463680.aspx.
[2] thuvienphapluat.vn. “Luật Công nghiệp Công nghệ số 2025.” September 11, 2025. https://thuvienphapluat.vn/van-ban/Cong-nghe-thong-tin/Luat-Cong-nghiep-cong-nghe-so-2025-so-71-2025-QH15-621341.aspx.
[3] thuvienphapluat.vn. “Nghị quyết 57-NQ/TW năm 2024 về Đột phá phát triển khoa học, công nghệ, đổi mới sáng tạo và chuyển đổi số quốc gia do Ban Chấp hành Trung ương Đảng ban hành.” May 12, 2025. https://thuvienphapluat.vn/van-ban/Cong-nghe-thong-tin/Nghi-quyet-57-NQ-TW-2024-dot-pha-phat-trien-khoa-hoc-cong-nghe-doi-moi-sang-tao-637245.aspx?anchor=dieu_2.
[4] VNN. “Vietnam Launches AI Ethics Committee to Lead Responsible Innovation.” Mst.gov.vn. https://english.mic.gov.vn, December 6, 2024. https://english.mst.gov.vn/vietnam-launches-ai-ethics-committee-to-lead-responsible-innovation-197241206162304766.htm.
[5] Oxford Insights. “Government AI Readiness Index.” Oxford Insights, 2024. https://oxfordinsights.com/ai-readiness/ai-readiness-index/.
[6] “Dự thảo Luật Trí tuệ nhân tạo,” 2025. https://mic.mediacdn.vn/document/2025/10/2/250925duthao-luatai-v10-1759393446665893284209.pdf. (in Vietnamese)
[7] European Commission. “AI Act.” European Commission, August 1, 2025. https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai.
[8] Nguyên Bảo. “Bộ trưởng Nguyễn Mạnh Hùng: Vấn đề của AI tạo ra có thể lớn hơn bom nguyên tử.” TUOI TRE ONLINE. tuoitre.vn, September 15, 2025. https://tuoitre.vn/bo-truong-nguyen-manh-hung-van-de-cua-ai-tao-ra-co-the-lon-hon-bom-nguyen-tu-20250915103427851.htm. (in Vietnamese)
