How Vietnam Brings AI into Practice: From Legislation to Real Solutions
Hanoi, vrijdag, 21 november 2025.
Vietnam is going beyond mere legislation: while the country presents an ambitious Artificial Intelligence (AI) bill with a focus on societal impact, companies such as Viettel Solutions are already implementing AI to address real pain points. Rather than technological showpieces, the focus is on tangible improvements—reducing internal errors by 30 to 50%, and deploying an electronic medical record system within two months, a timeline that would typically take months for a traditional rollout. The core principle? Use AI to solve real problems, not to impress. This combination of responsible regulation and practical application offers a valuable lesson for other nations, including the Netherlands, on how technology can genuinely contribute to everyday life.
Vietnam’s AI Legislation: Balancing Innovation and Social Responsibility
On Friday, 21 November 2025, Vietnam’s National Assembly will present the comprehensive draft law on artificial intelligence (AI) during its 10th session, a framework law consisting of eight chapters and 63 articles [2]. Proposed by Minister Nguyen Manh Hung, the project aims to institutionalize state and party policies, promote innovation, and manage risks, while simultaneously protecting national interests, human rights, and digital sovereignty [2]. The core principle of the law is: ‘artificial intelligence should serve humanity, not replace it’ [2]. The legislation regulates AI based on risk levels, encourages domestic AI development and autonomy, and establishes systematic measures for the management, research, development, application, and deployment of AI systems, including foreign organizations using AI in Vietnam [2]. The Commission for Science, Technology and Environment, chaired by Nguyen Thanh Hai, emphasizes that AI development profoundly impacts employment, privacy rights, and cultural values, and recommends introducing principles for the quality of AI data—accurate, sufficient, clean, and dynamic—and mandatory principles for network and data security [2]. The law supersedes and repeals provisions from the Law on Digital Technology Industry No. 71/2025/QH15, thereby closing legal gaps [2]. The draft is a flexible framework law, adaptable to rapid technological advancements, and must remain consistent with other laws such as the Civil Code and the Law on Product and Goods Quality [2]. The Commission also advises adding a dedicated section on the use of AI in training, research, and innovation, as the current draft lacks regulations in this area [2]. The law plans to prohibit actions such as using AI to manipulate elections, generate deepfakes, commit fraud, or compromise national security [2]. The draft will be adjusted to accommodate the fast-paced evolution of AI and must remain consistent with laws including the Civil Code, the Law on Product and Goods Quality, and the Hanoi Convention (signed in October 2025) [2].
Open-Source Data for AI: The Foundation of a National AI Ecosystem
Vietnam is building a strategic AI infrastructure by launching the ViGen project, an open-source Vietnamese dataset designed for research, development, and application of artificial intelligence [5]. Initiated by the National Institute of Information (NIC), established under Decision No. 1131/QD-TTg dated 12 June 2025, and supported by Resolution 57-NQ/TW dated 22 December 2024 [5], the ViGen project is part of the Vietnam Innovation Challenge Program 2025 and is being implemented in collaboration with companies such as Meta Group, AI for Vietnam Organization, NVIDIA, Viettel AI, Misa, and Genetica [5]. The dataset aims to reflect Vietnamese linguistic identity, including natural expressions and cultural values, and is essential for developing a large-scale Vietnamese language model by 2030 [5]. The first version of the ViGen dataset is expected by October 2025, but as of 21 November 2025, it has not yet been publicly released, and its rollout status remains unclear [5]. The dataset will serve as the foundation for ‘Made in Vietnam’ AI solutions, targeting improvements in quality of life and the resolution of socio-economic challenges in areas such as education, healthcare, and security [5]. Data collection infrastructure is under development, including the definition of data standards and the construction of technical platforms for quality assurance, information security, and scalability [5]. The NIC director called for collaboration among experts, scientists, businesses, and organizations during the consultation meeting held on 19 November 2025 [5]. Establishing an open-source Vietnamese data infrastructure is seen as a strategic step laying the groundwork for a domestic AI ecosystem, which not only enhances the competitiveness of local research institutions and companies but also aligns with Vietnam’s open data policies and legislation [5].
Viettel Solutions: AI as a Tool for Real Pain Points, Not a Showcase
Viettel Solutions, a division of the Viettel Group, focuses on integrating AI into business processes and product development to deliver concrete, practical digital solutions [6]. Rather than focusing on technological showpieces, the company targets ‘pain points’ in internal operations—such as managing business processes and customer service—resulting in a 30 to 50% reduction in operational violations [6]. Viettel Solutions’ internal AI assistant accelerates document retrieval: employees now need only type a question, and answers appear within seconds [6]. The company invests in a hardware infrastructure equipped with thousands of GPUs for training and deploying large-scale AI models and aims to build a team of hundreds of specialized AI engineers capable of managing the entire AI lifecycle [6]. Viettel Solutions collaborates with Big Tech companies such as NVIDIA and Qualcomm to gain access to advanced technologies, even in laboratory stages [6]. In Hanoi, a pilot project used AI and telecom data to create a highly accurate traffic map, comparable to Google Maps but tailored to Hanoi’s urban characteristics [6]. Another success is the implementation of electronic patient records (EPRs) across 27 military hospitals, completed in approximately two months—an amount of time typically required for a single small hospital [6]. Viettel aims for ‘universal AI’—broadly applicable, fast, and secure—rather than ‘demonstration products’ or short-term breakthroughs [6]. The core philosophy of Viettel Solutions is that ‘AI is not a competitor but a tool’ that promotes automation and eliminates repetitive manual tasks, allowing people to focus on creative, complex, and critical thinking [6]. Le Thanh Cong, Deputy General Director of Viettel Solutions, emphasizes: ‘The key is choosing the right problems and focusing on those that are truly painful, because that is where technology can create the most tangible change’ [6]. The company has identified AI as a cornerstone of its strategy for the period 2026–2030 [6].
AI in the Public Sector: From Chatbots to AI-Driven Information Services
In the public sector, Vietnam is focusing on creating responsible AI applications that improve information accessibility and refine public service delivery. The Vietnamese AI draft law identifies the use of AI in training, research, and innovation as essential and includes a dedicated section for these domains, as the current draft lacks regulation in this area [2]. Although concrete use of AI in public information and outreach is not yet detailed in the sources, Amsterdam’s model offers a valuable international example: the City of Amsterdam has presented two prototypes under the AI Futures project, where AI works in the background to support citizens, or where a personal AI agent communicates with the municipality on behalf of the citizen [4]. These prototypes are technically feasible but raise critical questions about societal desirability, the balance between convenience and control, and ethical design [4]. As of 21 November 2025, it remains unclear whether Vietnam has launched a similar initiative in the public sector, but the draft law allows for AI applications in government, emphasizing accountability and human oversight for significant decisions [2]. In another context, a report by Vietnam’s Authority for Electronic Communications and Information Technology highlights the introduction of guidelines on developer and implementer responsibility, with penalties for violations ranging from 100,000 to 500,000,000 dong (approximately €18,000 to €90,000) [3]. These guidelines have been in effect since 20 November 2025 and apply to both public and private sectors [3]. The application of AI in information and communication could, in the long term, lead to personalized information services, where citizens receive contextually relevant information based on their situation, goals, and language proficiency, and where campaign effectiveness is measured through real-time feedback [4]. The implementation of such systems depends on the adoption of the draft law and the development of responsible AI data infrastructure like ViGen [5].
The Challenges of AI: Privacy, Inclusivity, and Reliability
The application of AI in public information and outreach brings not only benefits but also significant challenges, particularly concerning privacy, inclusivity, and reliability. The Vietnamese draft law acknowledges that AI development has a major impact on privacy rights, employment, and cultural values, and therefore recommends introducing mandatory principles for data and network security [2]. While the draft law provides a framework for accountability and transparency, it remains unclear how this will be applied at the level of individual AI applications in public information [2]. The deployment of AI in the public sector, exemplified by Amsterdam’s AI Futures prototypes, raises concerns about citizens’ control over their data and the risk of technological dependency [4]. In Vietnam, incidents of fraud involving deepfake videos of Dutch celebrities shared via TikTok have already been reported, highlighting the dangers of AI-generated misinformation and manipulation [2]. The draft law therefore prohibits actions such as using AI to manipulate votes, engage in political games, or compromise national security [2]. The reliability of AI systems depends on the quality of the data they are trained on, underscoring the importance of initiatives like ViGen: an open-source dataset that is accurate, sufficient, clean, and dynamic [5]. Without this foundation, AI systems may generate bias, errors, and unreliable outputs. Inclusivity is another challenge: if AI data do not represent all groups in society, applications may exclude or discriminate against certain communities [5]. While Vietnamese legislation and the progress of ViGen are positive steps, it remains unclear whether explicit measures are in place to ensure inclusivity in AI applications for public communication [2]. The future effectiveness of AI in public information will depend on achieving a balance between technological innovation, responsibility, and ethical considerations.