Draft AI Law of Vietnam and impact on enterprises

Vietnam is becoming one of the first countries to enact a comprehensive legal framework to adjust the development, deployment and management of AI. The draft AI Law reflects Vietnam’s strategic intentions to assert digital sovereignty, improve technological competitiveness and attract high-quality investment in innovation ecosystems.

Unlike many other countries that are still in the exploration phase, Vietnam has been directly aiming to systematize the principles of AI governance, balancing the balance of risk control.

Scope, principles of direction and governance structure

The draft AI Law establishes a broad management range, beyond the scope of domestic developers, including any organization, whether domestic or foreign, with an AI system that impacts users, markets or national interests in Vietnam. The scope of application beyond the territory ensures that foreign technology providers cannot evade obligations by operating from abroad, which helps Vietnam catch up with the global management trends outlined in the EU’s AI Act.

According to the draft law, AI-related activities must comply with 7 basic principles that determine the national governance method, including: human-centered development; safety, fairness, transparency and accountability; national autonomy and international integration; comprehensive and sustainable development; balanced and harmonious policy planning; risk-based management; promoting EMITH. These principles lay out the legal framework, emphasizing responsible deployment without inhibiting technological progress.

To monitor the implementation, the draft Law introduces a centralized governance structure led by the Ministry of Science and Technology and a new national AI Commission.

The commission shall coordinate the implementation of regulations, manage the national AI database, approve high-risk AI operations and develop technical standards. The role of the committee is designed to ensure uniform oversight, reduce fragmentation and facilitate unifying policy between government, business and international partners.

Classification based on risk and obligation obligations

The draft AI Law applies a 4-level risk classification system, determining the level of control prescribed for AI systems. This structure helps Vietnam align with leading global management models, including the EU AI Act, but also adds tailored features to suit national security and socio-cultural considerations.

The four risk levels for evaluating AI systems, including:

Risks are unacceptable: AI systems are considered to be a threat to national security, human dignity or social order are strictly prohibited. These systems include cognitive behavioral manipulation technologies that deploy facial recognition on a large scale without the consent of users or creating deepfake videos that can destabilize public opinion.

High risk: The deployment of AI in areas such as finance, health, education, infrastructure and judiciary must be approved before being put on the market. These systems must undergo a mandatory government assessment and registration process before deployment.

Average risk: Systems that interact directly with users or create content must ensure user transparency and awareness. The supplier must clearly label the AI-generated outputs and deploy user feedback and monitoring mechanisms.

Low risk: AI systems with minimal social or economic impact are allowed to self-regulate, adhere to general principles and supervise after marketing.

Enterprises developing or implementing AI in Vietnam must conduct risk assessments and ensure compliance with detailed technical documents. High-risk AI systems must undergo compliance assessments, registrations in the national AI database and deploy human monitoring mechanisms and report incidents. Foreign suppliers must appoint a local legal representative to coordinate compliance and act as a link with Vietnamese authorities.

The draft law prohibits AI applications that pose systemic threats, including real-time biometric monitoring in public places (without special approval), large-scale facial recognition databases built through unauthorized collection of data and AI systems designed to manipulate public opinion or behavior in phishing manners.

Vietnam’s approach reflects the EU’s AI Act in using a risk-based classification model. However, Vietnam is more stressed on national sovereignty, data autonomy and infrastructure, as well as cultural stability.

While the EU prioritizes the harmony between member states, Vietnam’s framework is designed to promote the adoption of domestic AI, while maintaining the government’s supervision of strategic technologies. This demonstrates Vietnam’s commitment to the deployment of responsible AI, while providing specific routes to comply and enter the market.

National AI infrastructure, data management and innovative incentives

The draft AI Law lays the foundation for a national AI ecosystem, designed to promote innovation, while ensuring the transparency and supervision of the State. One highlight is the establishment of a national AI database, which serves as a foundation for registration and monitoring of AI systems operating in Vietnam.

High-risk AI systems must be registered before deployment, allowing authorities to monitor development activities, assess relevance and intervene as needed. High-risk system registration institutions and data contributions may enjoy incentives, such as access to infrastructure or financial support, as set out in the draft.

To mobilize resources, the Government encourages public-private partnerships and introduces a pioneering clause, recognizing AI models, algorithms and data assets as legal capital contributions. This represents a significant shift in management thinking, providing companies with new financial channels and supporting Vietnam’s digital economy’s ambition to formalize the digital economy.

The draft Law also proposes to establish AI clusters - Mekong areas, providing enterprises with the right to access common infrastructure, tax incentives, land use benefits and research facilities supported by the Government.

In addition to this, the National AI Development Fund will provide preferential funding, loans and funding to domestic startups, small and medium-sized enterprises and foreign investors developing AI capacity in Vietnam.

The regulatory framework testing mechanism allows companies to test AI applications under the legal conditions of being loosened, reducing barriers to entering the market and supporting responsible testing. These mechanisms emphasize Vietnam’s ambition to become an AI hub with clear incentives for industry participation.

Business impact and industry impact

The draft AI Law provides significant impacts for both domestic enterprises and foreign investors. Companies that develop or deploy high-risk AI systems must establish a strict compliance framework, including data security protocols, impact assessments and human monitoring mechanisms.

Foreign suppliers must appoint a legal representative in Vietnam, creating a clear point of responsibility for the regulatory authorities. Penalties can be fixed in the percentage of global revenue, enhancing strong cross-border determinism.

At the same time, this law creates strategic opportunities in key areas. In the field of healthcare, the application of AI in patient diagnosis and management will benefit from prioritizing access to infrastructure and national funding.

In the financial and financial technology sectors, AI can accelerate credit grading, fraud and digital payments in a tightly managed framework, helping to strengthen consumer confidence.

In the field of manufacturing and logistics, AI-based automation is in line with Vietnam’s industrial upgrade agenda, supporting supply chain efficiency and export competitiveness. Meanwhile, public services, including e-government platforms and smart city initiatives, will be supported through public-private partnerships and AI clusters.

Localization and integration R&D are currently deciding the advantage of entering the market, especially in areas such as healthcare and manufacturing. Companies soon engage in sandbox initiatives and clusters that can convert investment compliance into a pioneering advantage. This change allows businesses to use compliance as a strategic lever to expand the market.

Strategic prospects and recommendations for investors

The draft AI Law marks a restructuring in Vietnam’s digital policy, bringing AI from an unmanaged field into a strategic industry-oriented state-oriented.

The draft AI Law demonstrates a future-oriented approach, a balance between national security, ethical standards and economic competitiveness. Instead of limiting AI-driven industries, the law aims to create a controlled environment that encourages responsible innovation, investment trust and sustainable technological growth.

The schedule of deployment in each stage will be as follows:

From January 1, 2026: Establishment of legal infrastructure and initial implementation framework.

From July 1, 27, 2027, Fully apply its obligations to high-risk AI systems.

From July 1, 2029: Old high-risk AI systems begin transition phases, with 24 months allowed to register and implement compliance procedures.

Investors should use this timeline to evaluate AI assets, classify the level of risk and adjust the operating models to suit the upcoming compliance requirements. Pioneers shall benefit from policy incentives, priority access to national AI infrastructure and the opportunity to shape legal results through the participation of relevant parties.

Investors should prioritize 3 main actions: Integrate compliance in the product development process; participate in test programs (sandbox) to capture legal information and establish local partnerships or R&D activities to be eligible for incentives. By adjusting AI strategies in line with Vietnam’s national priorities, enterprises can turn regulatory compliance into competitive advantage. /.

Translated by Vietnam Journal of Science and Technology (VJST - MOST).