How is the high-risk artificial intelligence system managed from May 2026?

Legal Consulting Office of Lao Dong Newspaper answers:

Article 15, Decree 142/2026/ND-CP detailing a number of articles and measures to implement the Law on Artificial Intelligence (effective from May 1, 2026), stipulates the management of high-risk artificial intelligence systems as follows:

1. High-risk artificial intelligence system providers must establish and maintain a risk management system for the system they provide in accordance with the purpose of use, scope of deployment and level of risk of the system.

2. The risk management system specified in Clause 1 of this Article includes at least the following contents:

a) Identify and assess risks that may arise to human rights, safety, security or public interests in the process of designing, developing and providing the system;

b) Ensure the quality, suitability and representativeness of training data, test data and evaluation data within the necessary scope to limit risks arising from data;

c) Design and maintain human monitoring and intervention mechanisms suitable to the level of risk of the system;

d) Applying technical measures or management measures to prevent, limit or control identified risks;

e) Review and update risk management measures when the system has significant changes in model, data, operating method or purpose of use.

3. The Supplier is responsible for providing the deployment party with necessary information about the purpose of use of the system, safe operating conditions, identified risks and corresponding risk management measures to ensure the deployment and use of the system for the right purpose.

4. The developer of the high-risk artificial intelligence system must organize risk management during the deployment and operation of the system in accordance with the purpose of use, scope of deployment, level of risk of the system and technical guidance of the provider.

5. In the process of deployment and operation, the implementing party shall implement the following risk management measures:

a) Operate the system for the right purpose of use, usage conditions and usage limits determined by the supplier;

b) Organize monitoring of system operations to detect deviations, risks or incidents that may arise;

c) Establish and maintain human supervision and intervention mechanisms in the process of using the system in accordance with the law;

d) Applying measures to limit or control risks within the scope of its control when detecting systems operating for improper purposes or new risks arising;

e) Coordinate with suppliers in reviewing and updating risk management measures when the system has significant changes or when incidents occur.

6. In case a system is detected with the risk of causing serious damage to human life, health, human rights, property, network security, social order, social safety or public interests, the implementing party must promptly apply risk reduction measures within their control and notify the provider and competent state agency in accordance with the law.

Thus, from May 1, 2026, high-risk artificial intelligence systems are managed according to the above regulations.