Vietnam’s National Assembly passes AI Law requiring labels on AI-generated products

Vietnam’s National Assembly has passed the Law on Artificial Intelligence, setting out a requirement to label products created by artificial intelligence. The document raises the question of state management for a field being deployed rapidly in enterprises and the public sector, while establishing the responsibility of the party providing AI systems.

The content set out covers three axes: labelling AI products to distinguish machine-generated content, risk management according to the level of each type of system, and the responsibility of organisations providing AI systems. The risk-based approach allows obligations to be tiered instead of applying one identical set of requirements to every product.

For enterprises, the labelling requirement means content creation and digital product processes need to be reviewed, especially at stages where text- or image-generating models take part. Users gain an additional basis for recognising AI-generated content. Technical teams need to keep traceable records of the origin and of the way the system took part in each output product.

What needs noting is that the new law only sets a framework of principles; practical application depends on guiding documents and on how risk levels are classified for each group of systems. Organisations running AI systems should follow the process of concretising the rules and prepare dossiers describing the system, the scope of use and the control mechanisms.

Source: tuoitre.vn