Vietnam to require labels for realistic AI-generated content

Vietnam will introduce disclosure requirements for some AI-generated and AI-edited content from May 2026.

Under Decree 142/2026/ND-CP, organisations and individuals using AI systems must label content when AI has created or altered it in ways that could affect how users judge its authenticity.

The rules apply to audio, images, and videos. They are especially relevant to content that imitates the appearance or voice of real people, or recreates real-life events in a convincing way.

Disclosures must be clear, visible, and recognisable before or during access to the content. This means users should be able to identify that AI was involved before they treat the material as real or unaltered.

The decree targets a practical transparency problem. AI tools can now produce synthetic voices, realistic images, and edited videos that are difficult for ordinary users to distinguish from authentic recordings.

The requirement does not appear to label all AI use. Its focus is narrower: content where AI generation or editing could mislead users about whether a person actually appeared, spoke, or whether an event really happened as shown.

For platforms, media producers, advertisers, and individual creators, the rule creates a new compliance obligation around presentation, not only content production. The key question will be whether labels are placed early enough and clearly enough for users to understand what they are seeing or hearing.