Vietnam

During an inspection or examination, the organizations and individuals concerned must provide technical dossiers, trace logs, training data and other necessary information to determine the cause of a violation or incident or to apportion responsibility. That information must be provided in compliance with the law on protecting state secrets, data, personal data and intellectual property.

The agencies and organizations assigned state management of artificial intelligence are responsible for checking that organizations and individuals comply with the law in their artificial intelligence activity.

Circular No. 05/2026/TT-BKHCN of the Ministry of Science and Technology issues the National Artificial Intelligence Ethics Framework, which applies to state management agencies and to any organization or individual that develops, provides, deploys or uses an AI system for state management or public services.

Under it those actors are responsible for identifying harmful scenarios in advance and building preventive measures, for setting quality criteria for data, models and outputs and building internal testing, validation and verification before deployment, and for building human oversight and intervention mechanisms proportionate to the system's impact. Other Vietnamese and foreign organizations and individuals working with AI in Vietnam are encouraged, not required, to apply it. The Circular took effect on 10 March 2026.

Article 7 bars developing, providing, deploying, or using an AI system to use forged elements or simulations of real people or events to intentionally and systematically deceive or manipulate human perceptions or behaviour where this causes serious harm, to exploit weaknesses of a vulnerable group such as children, the elderly, people with disabilities, or people with limited cognitive or civil-act capacity, or to create or disseminate fake content capable of posing a serious danger to national security, public order, or social safety.

It separately bars collecting, processing, or using data to develop, train, test, or operate an AI system in a way that contravenes Vietnam's data, personal data protection, intellectual property, or cybersecurity law, and bars obstructing, disabling, or distorting the human-oversight mechanisms the Law requires over an AI system.

A provider must classify its AI system itself before putting it into use, and answers before the law for the accuracy and truthfulness of the result. A system that is not classified as high risk but falls within Article 9 of the Decree is classified as medium risk. Article 9 reaches a system that can confuse, influence or manipulate users because they cannot tell that they are dealing with an AI system or with content it made.

A deployer must work with the provider to re-classify a system when its own modification, integration or change of function or purpose creates a new or higher risk. When re-classification finds a higher risk level, the provider or deployer must notify the competent state agency within 15 working days of finishing the review and apply the matching risk-management measures at once.

A provider must prepare a risk classification dossier for a high-risk or medium-risk system before putting it into use. A provider or deployer must keep the classification dossier for as long as the system operates. A provider of a medium-risk or high-risk system must notify the classification result to the Ministry of Science and Technology through the one-stop AI portal before putting the system into use.

A provider of a high-risk system must complete a conformity assessment before putting it into use. The provider must repeat the assessment when a significant change affects the original result. A system on the list of high-risk systems that need certification before use must be assessed through a conformity assessment organization. A provider that assesses a high-risk system itself must prepare a technical dossier and answers before the law for the result.

A provider must publish the conformity assessment result on the one-stop AI portal before putting the system into use and when a re-assessment updates it. A provider of a high-risk system must set up and maintain a risk-management system suited to the system's purpose, scope and risk level. The provider must design and maintain a mechanism for human oversight and intervention suited to the system's risk level.

A provider must give the deployer the information it needs on the system's purpose, safe operating conditions, identified risks and matching risk measures. A deployer of a high-risk system must operate it within the purpose, conditions and limits of use that the provider sets. A deployer that finds a system at risk of serious harm must promptly limit the risk within its control and notify the provider and the competent state agency. The Decree takes effect on 1 May 2026.

A serious incident includes an event in the operation of an AI system that causes loss of life or serious harm to a person's health. A deployer and a user must promptly record a serious incident, take the measures needed to limit its consequences and notify the provider. A provider must apply the technical measures needed to prevent, limit and remedy the consequences of the incident.

A provider or deployer must file a preliminary report through the one-stop AI portal within 72 hours of confirming an urgent serious incident, meaning one under points a and d of Article 19(1) or an uncontrollable one under point c. For any other serious incident the preliminary report is due within five working days of confirmation.

The time of confirmation runs from when the reporting person has enough initial information to establish that the incident really happened and probably stems from a fault of the AI system, without waiting for a full investigation of the technical cause. A deployer that cannot reach the provider must make the report itself. A provider and a deployer must keep the system logs, data and information on the incident for verification, assessment and remedy.

They must send the competent state agency a formal report on the remedy within 15 days of the preliminary report. The Decree takes effect on 1 May 2026.

Article 12 requires every developer, provider, deployer and user of an artificial intelligence system to keep the system safe, secure and reliable, and to promptly detect and remedy an incident capable of harming people, property, data or social order.

When a serious incident occurs in an artificial intelligence system, the developer and provider must urgently apply technical measures to remedy, suspend or recall the system, and must simultaneously notify the competent state agency of the incident. The deployer and the user of the system must record the incident, notify it promptly, and coordinate with one another during the remediation process.

The competent state management agency receives, verifies and guides the remediation, and may require the system to be suspended, recalled or reassessed when necessary. Reporting and remediation of the incident is carried out through the one-stop artificial intelligence portal.

Article 12 does not itself state a reporting deadline; instead it directs the Government to specify the reporting procedure and the responsibilities of the agencies, organisations and individuals involved, calibrated to the severity of the incident and the scope of its impact. Article 12 addresses the developer, provider, deployer and user of the system and the competent state agency; it does not itself direct any notice to a person affected by the incident.

Decree 142/2026/ND-CP sets the preliminary-report deadline for an urgent serious incident at 72 hours from confirmation. For any other serious incident the Decree makes the preliminary report due within five working days of confirmation.

Article 9 classifies an AI system as high, medium, or low risk based on its potential to cause significant harm to rights, safety, security, or public interest, its capacity to confuse users about whether they are dealing with an AI system, and criteria the Government is to detail further, including the field of use and the scale and range of its impact.

Article 10 requires a provider to classify its own system before putting it into service, to prepare a classification dossier for a medium-risk or high-risk system, and to notify the classification result to the Ministry of Science and Technology through the one-stop AI portal before deployment; a deployer inheriting a provider's classification must re-classify if a modification creates a new or higher risk.

Articles 13 and 14 require a high-risk system to undergo conformity assessment, by a registered assessment organization or by the provider itself depending on the system, before being put into service or after a significant change, and require the assessment result to be maintained throughout operation as a condition of continued use.

Decision No. 367/QD-TTg, the government's implementation plan for the Law, is preparing a Prime Ministerial decision listing high-risk AI systems by five criteria, but this list had not been issued as of this review.

A provider and a deployer must give information on the system's purpose of use, scope, conditions of use and limits. They must keep the information and documents needed for inspection and supervision. They must explain themselves when the competent state agency asks in writing. A provider must apply a technical solution so that audio, image or video output is marked in a machine-readable format. A provider need not apply machine-readable marking to text output unless another law requires it.

The provider must keep the marking function working while content is created, exported and supplied within the functions the system controls. A provider of an open-source or free system meets the marking duty by building in the marking function or by publishing the tools, standard configuration, interface or technical documents that let a deployer set it up. A deployer that uses such a system to supply content to the public must apply a technical solution to mark the output.

A deployer must give clear notice when it makes available to the public AI-generated or AI-edited content that could confuse people about the authenticity of an event or a person or about the origin of the content. A deployer must attach an easily recognizable label to AI-made or AI-edited audio, images or video that simulate a real person's appearance or voice or recreate a real event, to tell them apart from real content.

The notice or label must be given before or at the time the recipient reaches the content. It must not be designed to hide or weaken recognition of the content's nature. A deployer need not label content that is used only inside the organization and is not made available to the public. The Decree takes effect on 1 May 2026.

Article 11 requires a provider to design and operate an AI system that interacts directly with humans so that a user can recognise they are interacting with such a system, unless the law provides otherwise, and to mark AI-generated audio, image, and video content in a machine-readable format under forthcoming Government regulations.

A deployer must clearly notify the public when providing AI-generated or AI-edited text, audio, images, or video that could cause confusion about the authenticity of an event or a person, and must attach an easily recognisable label to AI-generated or AI-edited content that simulates or replicates a real person's appearance or voice or recreates a real event, with a lighter labeling standard for cinematographic, artistic, or creative works so the label does not obstruct enjoyment of the work.

Both providers and deployers must keep this transparency information current throughout the life of the system, product, or content.

Article 31 defines biometric data as data on physical attributes and unique and stable biological characteristics of a person used to identify that person. It requires an organization, agency, or individual collecting or processing biometric data to apply physical security measures for storage and transmission devices, limit access, maintain a monitoring system to detect infringement, and comply with relevant laws and international standards.

Where the processing damages the data subject, the collecting and processing party must notify that data subject as the Government prescribes. Article 31 names no facial, voice, or fingerprint example, resting entirely on the general physical-attribute language; whether the definition reaches an identifier derived from a photo, video, or audio recording is unresolved.

This provision functions as heightened, category-wide protection rather than a named consent-at-capture rule distinct from the Law's general Article 9 consent duty.

Article 23 requires the personal data controller, the personal data controlling and processing party, or a third party to notify the agency in charge of personal data protection within 72 hours after detecting a violation likely to cause harm to national defense and security and social order and safety, or to infringe upon the life, health, honor, dignity, or property of the personal data subject.

Where a personal data processor detects a violation, it must promptly notify the controller or the controlling and processing party.

The Law on Personal Data Protection (Law No. 91/2025/QH15) is Vietnam's first standalone, comprehensive personal-data statute. Decree 13/2023/ND-CP, which was itself only a government decree rather than a National Assembly law, preceded it, and Article 39 lets consents given under that Decree continue without being obtained again. Article 38(1) fixes its commencement at 1 January 2026. The signed text records its passage by the 15th National Assembly at its 9th session on 26 June 2025.

A companion implementing decree, Decree 356/2025/ND-CP, issued on 31 December 2025, took effect alongside the Law on 1 January 2026. A citation of this same enactment as Law No. 34/2025/QH15 is incorrect, because the signed text is numbered 91/2025/QH15.

Article 20(1) lists the cross-border transfers the Law covers: moving personal data stored in Vietnam to a storage system outside Vietnam, an agency, organization, or individual in Vietnam passing personal data to a counterpart abroad, and using a platform outside Vietnam to process personal data collected in Vietnam.

Under Article 20(2), whoever transfers personal data across the border must prepare a cross-border transfer impact assessment dossier and send one original to the specialized personal data protection agency within 60 days of the first day of transfer. Under Article 20(3), the assessment is done once for the whole period of the party's operation and updated as Article 22 requires.

Under Article 20(5), that agency decides to require a transfer to stop when the data is being sent for use in activity that may harm national defense or security. Article 20(6) excuses transfers by competent state bodies, an organization storing its own employees' data on a cloud service, and a data subject transferring his or her own data. The paragraphs of Article 20 set no adequacy list, no approval step, and no data-localization element.

Article 20(7) leaves the detailed conditions and procedures to the Government. A separate penalties article sets the maximum fine for an organization violating the cross-border transfer rules at 5 percent of its prior-year revenue, a tier distinct from the Law's general fine.

Article 33(1)(a) names the specialized personal data protection agency under the Ministry of Public Security as part of the personal data protection force. Article 36(2) makes the Ministry of Public Security the focal agency answerable to the Government for state management of personal data protection, except for matters within the Ministry of National Defence's scope.

Under Article 8(4), the administrative fine for an organization violating the cross-border transfer rules reaches up to 5 percent of the prior year's revenue. Under Article 8(3), the fine for buying or selling personal data reaches up to 10 times the proceeds. Under Article 8(5), the fine for other violations reaches up to VND 3 billion. Under Article 8(6), individuals are capped at half the organizational maximum.

Article 4(1) gives individuals the right to request the provision, deletion, or restriction of processing of their personal data and to object to its processing. Article 4(1) also gives individuals the right to file complaints and denunciations, initiate lawsuits, and request compensation for damage in accordance with law, arming a private plaintiff.

Article 3(2) requires an institution to categorize its information systems in the order prescribed at article 5 and to apply the suitable information security policy. Article 42(2) requires an institution to assess the security of Level 3 and above information systems and other information systems holding client personal information before use.

Article 42(3) sets the assessment interval at six months for Level 5 systems, one year for Level 4 and 3 systems and devices connecting to the Internet, clients or third parties, and two years for the whole institution. Article 42(4) requires assessment results to be reported in writing to the legal representative of the institution and the competent authority.

Article 1(2) applies the Circular to credit institutions, branches of foreign banks, intermediary payment service providers, credit information companies, the National Payment Corporation of Vietnam, the Vietnam Asset Management Company, the National Banknote Printing Plant and Deposit Insurance of Vietnam. Article 55(3) makes the banking investigation and supervision agencies responsible for handling administrative violations of the Circular as per legal regulations. The Circular takes effect on January 01, 2021.

Article 9(1) of the Law on Cybersecurity defines information systems critical to national security as those with a strategic and especially important role whose incidents or violations of the law on cybersecurity may cause damage to national security or serious damage to social order and safety, that are on the list issued by the Prime Minister.

Article 9(2)(g) includes national information systems in the energy, finance, banking, telecommunications, transport, agriculture, natural resources and environment, chemicals, healthcare and culture sectors among them. Article 9(3) requires these systems to undergo cybersecurity appraisal and be certified as meeting cybersecurity conditions before being put into operation, to be regularly inspected and monitored during operation, and to respond to and remedy cybersecurity incidents promptly.

Article 11(1)(b) requires the manager of such a system to conduct a cybersecurity inspection before putting a new or upgraded system into operation and to notify the results of its annual cybersecurity self-inspection in writing to the competent specialised cybersecurity protection forces before October every year.

Article 11(1)(d) requires the manager to develop plans for responding to and remedying cybersecurity incidents, to implement them when incidents occur, and to report promptly to the competent specialised cybersecurity protection forces, with no fixed number of hours stated. Article 11(2) gives the Ministry of Public Security the tasks of appraising, certifying and inspecting these systems, other than military and cryptographic information systems. The Law takes effect on July 1, 2026.

Article 8 sorts every information system into one of five levels by the damage a cybersecurity incident or violation could cause, and Article 10 grades a manager's protection duty to that level, reaching any manager, state or private, of a level 1 to level 4 system; a system on the Prime Minister's list of information systems critical to national security carries a fuller, separately gated regime under Articles 9 and 11.

Decree No. 331/2026/ND-CP, effective 19 August 2026, sets the classification criteria and the approval procedure the Law's Article 8(2) delegates to the Government.

Article 54(1) requires an institution to report a cyber security incident to the State Bank (Information Technology Authority) within 24 hours starting from the detection of the incident and within 5 working days after incident handling. Article 2(2) defines an information security incident as an incident in which digital information and an information system are attacked or harmed, resulting in negative effects on their confidentiality, integrity and availability.

Article 1(2) applies the Circular to credit institutions, branches of foreign banks, intermediary payment service providers, credit information companies, the National Payment Corporation of Vietnam, the Vietnam Asset Management Company, the National Banknote Printing Plant and Deposit Insurance of Vietnam. Article 55(3) makes the banking investigation and supervision agencies responsible for handling administrative violations of the Circular as per legal regulations. The Circular takes effect on January 01, 2021.

Article 40(1)(c) requires any information system manager, at any level, to report a cybersecurity incident to the specialised cybersecurity protection force.

Article 41 layers a parallel duty on any domestic or foreign enterprise providing a telecommunications, Internet, or other value-added cyberspace service in Vietnam: keep an emergency response plan for cybersecurity weaknesses, risks and incidents, implement it and report an incident the moment it occurs, and apply technical measures to secure data processing, including personal-data processing.

Article 54 requires an agency, organization, or individual that manages or provides information and communications products and services, or organizes activities, in the Internet environment to apply measures ensuring safety and privacy for children, defined by Article 1 as a person below the age of sixteen; clause 3 leaves the Government to detail the article's specific requirements, which are not independently confirmed here.

Article 87 assigns the Ministry of Information and Communications responsibility to manage and guide regulations on the press, publication, telecommunications, Internet, and other information channels involving children, and to protect children in the environment of the Internet, computer networks, and telecommunication networks.

Article 105 provides that a violator of the Law is disciplined, administratively sanctioned, or examined for penal liability depending on the nature and severity of the violation.