Vietnam's AI Law 2026: what should business owners do?

Who the law applies to

It covers Vietnamese and foreign organisations and individuals taking part in AI activity in Vietnam: research, development, provision, deployment and use of AI systems. AI used only for defence, security or cryptography is outside its scope.

A shop using ChatGPT or image and voice tools is a user. When it puts a chatbot in front of customers it is deploying an AI system for others, so the points below matter even more.

Labelling AI content

Under Article 11, audio, images and video created or edited with AI to imitate a real person's appearance or voice, or to recreate real events, must be clearly labelled; creative works are labelled in a fitting way. In a shop this usually means:

• AI-generated models wearing your clothes: write "AI-generated image" on the photo or the post.
• Videos with an AI voice: say so in the description and at the start of the video.
• Product photos edited by AI until they differ from the real item: label them, since customers can be misled.
• On YouTube, use the altered or synthetic content disclosure when a video looks real.

What is not allowed

Article 7 bans using AI to deceive, manipulate perception or create harmful fakes. For a shop: no AI faces or voices of real people, celebrities or customers in ads without consent, and no AI claims about products that are untrue.

Chatbots and customer data

Label the chatbot as an "AI assistant" at the start of the chat and always offer a way to reach a person. It is honest, and it keeps trust when the AI gets something wrong.

Messages, phone numbers and addresses you put into AI tools are personal data. Vietnam's personal data protection law has applied since 1 January 2026: share only what is needed, and pick services that say how data is used.

A to-do list for now

• List where the shop uses AI: images, voices, video, copy, chatbots.
• Label AI images, voices and video that imitate real people or could mislead.
• Mark the chatbot as an AI assistant and let customers reach a person.
• Review ads: no real faces or voices without consent.
• Read the data terms of every AI tool you use.
• Keep a note of how you label, in case customers or authorities ask.

Common questions

• When did the AI Law take effect?

  On 1 March 2026. Law No. 134/2025/QH15 was passed by the National Assembly on 10 December 2025; Decree 142/2026/ND-CP, which details it, applies from 1 May 2026.

• Do posts written with AI need a label?

  The labelling rule in Article 11 is about audio, images and video imitating real people or events. A post drafted by AI and checked by you usually falls outside it, but you are still responsible if it says something untrue about a product.

• If AI model photos are labelled, what risk is left?

  The photo still has to match the real item: colour, fit, material. A label tells customers the image is AI-made; it does not make a misleading product photo accurate.