Vietnam - AI Governance Law (2025)

Overview

The Law on Artificial Intelligence (Law No. 134/2025/QH15) represents a landmark legislative effort by Vietnam to establish a comprehensive legal framework for the development, application, and governance of artificial intelligence systems within the country. Passed by the National Assembly on December 10, 2025, and entering into force on March 1, 2026, this law aims to foster innovation in AI while simultaneously ensuring safety, transparency, and accountability across various sectors. The law's scope extends to all organizations and individuals, both Vietnamese and foreign, involved in the research, development, provision, deployment, and use of AI systems within Vietnam, excluding those exclusively for defense, state security, or intelligence purposes.

Definitions

The Law on Artificial Intelligence (2025) provides crucial definitions to delineate its scope and application, ensuring clarity for stakeholders involved in the AI ecosystem. Key terms such as "Artificial Intelligence System" are defined broadly to encompass a wide range of technologies capable of performing tasks that typically require human intelligence, including learning, reasoning, problem-solving, perception, and decision-making. This broad definition is intended to be technology-neutral and future-proof, accommodating evolving AI paradigms. The law distinguishes between different types of AI systems, particularly focusing on the classification of "High-Risk AI Systems." These are identified based on their potential to cause significant harm to health, safety, fundamental rights, or critical infrastructure, necessitating stricter regulatory oversight and compliance requirements.

Further definitions include "AI Provider," referring to any natural or legal person who develops an AI system or places an AI system on the market under their name or trademark, and "AI Deployer," which designates any natural or legal person who uses an AI system under their authority. The law also elaborates on concepts such as "Training Data," referring to the datasets used to train AI models, and emphasizes the importance of data quality, representativeness, and bias mitigation. The definition of "Human Oversight" is critical, detailing the necessity for human intervention and control over AI systems, especially those deemed high-risk, to ensure ethical outcomes and accountability. These definitions collectively form the foundational lexicon for implementing the law, guiding both innovators and regulators in their respective roles and responsibilities within Vietnam's burgeoning AI landscape.

Governance and Institutional Framework

The governance structure established by Vietnam's Law on Artificial Intelligence (2025) is designed to provide robust oversight and strategic direction for the country's AI development. While the specific institutional bodies are not exhaustively detailed in publicly available summaries, the law implicitly tasks the Ministry of Science and Technology (MOST) with a central role in state administration of AI activities, building upon its existing mandate for science and technology development, intellectual property, and standards. This ministry is expected to coordinate efforts across various government agencies to ensure a coherent national approach to AI. The legislation likely mandates the establishment or designation of specific committees or agencies responsible for implementing the law's provisions, including developing technical standards, conducting conformity assessments, and managing AI-related risks. The overarching goal is to create an autonomous AI ecosystem, enhancing national competitiveness and ensuring proactive opportunity seizure while managing associated risks.

This framework is also expected to facilitate public-private partnerships, encouraging collaboration between government, academia, and industry to drive AI research, development, and application. The law aims to provide a clear legal corridor for sustainable AI development, emphasizing a balance between strict regulation and the promotion of innovation. It is anticipated that the government will issue subordinate legislation, such as decrees and circulars, to detail the functions, tasks, powers, and organizational structures of the responsible entities, as well as specific guidelines for implementation. The framework is also likely to include mechanisms for international cooperation and alignment with global AI governance best practices, reflecting Vietnam's commitment to active integration with new technological standards. This multi-faceted approach aims to ensure that AI development in Vietnam is guided by clear policies, supported by appropriate institutions, and responsive to both national priorities and global trends.

Key Focus Areas

The Law on Artificial Intelligence (2025) in Vietnam delineates several key focus areas to ensure responsible and innovative AI development. A primary focus is on risk management, adopting a tiered approach where AI systems are classified based on their potential to cause harm. This classification dictates the level of regulatory scrutiny and the obligations placed on AI providers and deployers, including requirements for risk assessments, mitigation strategies, and post-market monitoring. The law aims to regulate the outputs and usage behaviors of AI systems rather than intervening in the underlying models, thereby ensuring that innovation is not hindered while still prioritizing safety and accountability. This balanced approach is critical for fostering a dynamic AI industry without compromising public trust or welfare. Provisions are expected to cover areas such as cybersecurity risks associated with AI systems and the need for robust data infrastructure to support secure and ethical AI development.

Another significant focus area is transparency and disclosure. The law is expected to mandate requirements for explainability, ensuring that users can understand the logic and factors behind AI-driven decisions, especially in high-stakes applications. This includes provisions for disclosing when users are interacting with an AI system and potentially labeling AI-generated content to prevent misinformation or manipulation. The emphasis on transparency is crucial for building trust in AI technologies and empowering individuals to make informed decisions. Furthermore, data protection and privacy are implicitly addressed, as AI systems heavily rely on data. While the AI law itself may not be the primary instrument for data protection, it is expected to interact closely with existing or forthcoming data privacy regulations, ensuring that AI development and deployment adhere to strict standards for personal data handling, consent, and data subject rights. The law also promotes the development of an autonomous AI ecosystem, including expanding access to shared data and supporting the testing of new models, alongside incentives for AI enterprises and human resource development.

Implementation Framework

The implementation framework for Vietnam's Law on Artificial Intelligence (2025) is envisioned as a multi-stage process, commencing with the law's effective date of March 1, 2026. This phased approach is crucial for allowing stakeholders sufficient time to understand and comply with the new regulations. Initially, the focus will be on establishing the necessary institutional infrastructure, including designating or creating the specific government bodies and committees responsible for overseeing AI governance. This will involve defining their mandates, staffing, and operational procedures to ensure effective enforcement and strategic guidance. A critical aspect of this phase will be the development and promulgation of subordinate legislation, such as decrees, circulars, and technical standards, which will provide detailed guidance on the practical application of the law's provisions. These secondary instruments are essential for translating the high-level principles of the law into actionable requirements for AI providers and deployers.

Subsequent stages of implementation will likely involve the development of national AI strategies and action plans, outlining specific initiatives to promote AI research, development, and application while ensuring compliance with the regulatory framework. This includes fostering public-private partnerships, investing in AI infrastructure, and supporting human resource development in the field. The government is expected to establish mechanisms for industry engagement, providing platforms for dialogue and collaboration to ensure that the regulatory environment remains responsive to technological advancements and industry needs. Furthermore, the implementation framework will encompass the development of compliance tools, such as guidelines, best practices, and conformity assessment procedures, to assist organizations in meeting their legal obligations. Special attention will be given to supporting small and medium-sized enterprises (SMEs) in adopting and complying with AI regulations, recognizing their vital role in the innovation ecosystem.

Monitoring and Evaluation

The Law on Artificial Intelligence (2025) in Vietnam incorporates robust mechanisms for monitoring and evaluation to ensure its effectiveness and adaptability in a rapidly evolving technological landscape. Continuous monitoring will be conducted by the designated regulatory bodies, likely under the Ministry of Science and Technology, to track the implementation of the law's provisions and assess its impact on AI development, innovation, and societal well-being. This includes overseeing compliance by AI providers and deployers, reviewing risk assessments, and evaluating the efficacy of mitigation measures for high-risk AI systems. The monitoring process will also involve collecting data on AI-related incidents, breaches, and complaints to identify emerging risks and areas requiring regulatory adjustment. Regular reports on the state of AI development and regulation in Vietnam are expected to be published, providing transparency and accountability to the public and international community.

Periodic evaluations of the law's overall effectiveness will be conducted to determine if its objectives are being met, particularly regarding the balance between fostering innovation and ensuring safety and ethical use of AI. These evaluations will involve comprehensive impact assessments, analyzing the economic, social, and ethical implications of AI technologies and the regulatory framework. Stakeholder consultations, including input from industry, academia, civil society, and international experts, will be integral to the evaluation process, ensuring a multi-perspective assessment. Based on the findings of these evaluations, the government will be empowered to propose amendments to the law or its subordinate legislation, ensuring that the regulatory framework remains relevant, proportionate, and future-proof. This iterative approach to monitoring and evaluation underscores Vietnam's commitment to adaptive governance in the dynamic field of artificial intelligence, allowing for necessary adjustments to maintain a competitive edge while upholding fundamental values.

Penalties, Liability, and Appeals

The Law on Artificial Intelligence (2025) establishes a clear framework for penalties, liability, and appeals to ensure compliance and provide redress for damages resulting from AI systems. The law stipulates that organizations and individuals found in violation of its provisions will face sanctions, which may include fines, suspension of service provision, or other administrative penalties. The severity of these penalties will likely be proportionate to the nature and gravity of the infringement, taking into account factors such as the level of risk posed by the AI system, the extent of harm caused, and whether the violation was intentional or negligent. Importantly, the law moves away from revenue-based fines, focusing instead on comprehensive compensation for all damages incurred due to AI system failures or misuse. This approach emphasizes restitution for affected parties and aims to deter irresponsible AI development and deployment.

Regarding liability, the law clarifies the responsibilities of AI providers and deployers for damages caused by their AI systems, particularly those classified as high-risk. While the law does not classify high-risk AI systems as inherently dangerous sources in the same vein as physical hazards, it acknowledges the risks arising from AI decision-making. This means that liability provisions will likely focus on failures in design, development, testing, deployment, or oversight that lead to adverse outcomes. The law is expected to outline procedures for individuals or entities to seek redress for damages, including mechanisms for filing complaints, initiating investigations, and pursuing compensation. Furthermore, provisions for appeals will be in place, allowing aggrieved parties to challenge regulatory decisions or dispute liability determinations. This comprehensive system of penalties, liability, and appeals is crucial for instilling confidence in AI technologies, protecting user rights, and ensuring that accountability is upheld throughout the AI value chain in Vietnam.

Relationship to Other Instruments

The Law on Artificial Intelligence (2025) in Vietnam is designed to operate within the existing legal landscape, establishing clear relationships with other pertinent legal instruments. It is expected to complement and interact with Vietnam's cybersecurity laws, ensuring that AI systems are developed and deployed with robust security measures to protect against cyber threats and data breaches. Given the data-intensive nature of AI, the law will also have a significant relationship with data protection and privacy regulations, ensuring that the collection, processing, and use of data for AI purposes comply with established privacy principles, consent requirements, and data subject rights. This synergistic approach aims to create a cohesive regulatory environment where AI innovation thrives without compromising fundamental rights or national security. The AI law is also expected to inherit and, where necessary, abolish AI-related regulations from other existing laws, such as the Law on Digital Technology Industry, to streamline and improve the legal system governing AI.

Furthermore, the Law on Artificial Intelligence will likely interact with intellectual property laws, particularly concerning the ownership and protection of AI-generated content, algorithms, and training data. It may also influence or be influenced by sector-specific regulations, such as those governing AI applications in healthcare, finance, or transportation, providing a general framework that can be tailored to the unique characteristics and risks of different industries. The law's emphasis on human oversight and ethical AI development also aligns with broader human rights frameworks and principles of non-discrimination. By establishing these clear relationships, the Law on Artificial Intelligence aims to avoid regulatory fragmentation and create a predictable legal environment that supports both technological advancement and responsible governance. This integrated approach ensures that Vietnam's AI strategy is not an isolated endeavor but rather a harmonized component of its broader digital transformation agenda.

International Alignment

Vietnam's Law on Artificial Intelligence (2025) demonstrates a strong commitment to international alignment, reflecting the country's ambition to integrate with global AI governance standards and best practices. During its drafting process, the government carefully studied international experiences, selectively referencing legislation and practical insights from various jurisdictions. This comparative approach allowed Vietnam to adopt a balanced regulatory model that seeks to harmonize safety and development, drawing lessons from stricter European regulations, development-focused Japanese approaches, and South Korea's combined model. The aim is to create a framework that is robust enough to ensure safety and accountability, yet flexible enough to foster innovation, positioning Vietnam competitively in the global AI landscape. This proactive engagement with international norms is crucial for facilitating cross-border cooperation, mutual recognition of AI systems, and the free flow of ethical AI innovation.

The law is expected to support Vietnam's active participation in international forums and initiatives related to AI governance, contributing to the development of global standards and norms. This includes aligning with principles promoted by international bodies such as the OECD, UNESCO, and the UN, particularly concerning ethical AI, human rights, and sustainable development goals. By ensuring its legal framework is compatible with international standards, Vietnam aims to enhance its attractiveness as a destination for foreign investment in AI and facilitate collaborations with international partners. The law's provisions on addressing issues like AI-generated content and the responsibilities of cross-border AI service platforms further underscore this commitment to international integration and the maintenance of digital sovereignty in a globally interconnected digital economy. This strategic alignment is vital for Vietnam to leverage global expertise, contribute to the responsible evolution of AI worldwide, and secure its position as a significant player in the international AI arena.

Implementation Timeline

Sources and References