Vietnam's Decree 147: Real Names, Takedowns, and AI | TLY

AI Regulation Tracker  /  Regulation in force

Vietnam's Decree 147 Forces Real-Name Verification and 24-Hour Takedowns on Platforms

Vietnam's internet decree, in force since December 2024, already binds platforms serving Vietnamese users to verify every account and pull flagged content fast. It set the identity-and-moderation groundwork a full year before the country's AI Law.

Vietnam's Decree 147 Forces Real-Name Verification and 24-Hour Takedowns on Platforms regulation briefing
The Leveraged Years AI Regulation Tracker

Vietnam's rulebook for the open internet changed at the end of 2024, and most firms serving Vietnamese users are already inside its reach. Decree No. 147/2024/ND-CP, issued on November 9, 2024 and in force from December 25, 2024, replaced Decree 72/2013 as the master regulation for internet services and online information. Its two hardest edges are identity and speed: platforms must know who their users are, and they must take flagged content down quickly.

What Decree 147 actually requires

The account-authentication rule is the provision with the widest operational blast radius. Regulated social networks, both onshore and offshore, must verify active user accounts against a Vietnamese mobile phone number or a national identification number. Providers had 90 days from the effective date, meaning until March 25, 2025, to authenticate their existing active users. Tilleke & Gibbins reads the text as binding both domestic operators and offshore platforms that meet the coverage tests.

The second edge is removal speed. Under the decree, providers must take down content, services, or applications that violate Vietnamese law within 24 hours of a request from a competent authority, and must resolve user complaints about content that breaches the Cybersecurity Law within a comparable short window. That clock is short enough that manual, case-by-case review will not keep pace, so covered platforms need a standing process rather than an ad hoc one. The decree also pulls firms into a notification regime: an offshore provider that leases onshore storage or draws more than 100,000 monthly visits from Vietnam over six consecutive months must formally notify the authorities within 60 days of crossing that line, which turns the thresholds into a monitoring task, not a one-time check.

The AI-labeling question

Decree 147 is frequently discussed alongside Vietnam's move to label AI-generated media, and the two are connected as policy, but they are not the same instrument. On a reading of the decree text and the leading secondary summaries, Decree 147's confirmed duties are identity verification and rapid content removal, not a standalone machine-readable label on AI audio, image, and video. The explicit AI-content labeling mandate sits in Vietnam's AI Law, No. 134/2025/QH15, and its implementing decree, which took effect in 2026 and requires providers to mark AI-generated media and requires deployers to warn users when synthetic content could be mistaken for real people or events. Decree 147 is best understood as the identity-and-moderation layer that landed a full year before those AI-specific rules, and that already governs how flagged synthetic content gets pulled.

What it does not do

The decree does not, on its own, set a technical standard for watermarking or a universal AI disclosure line. It does not create a private right of action, and it does not replace the sector rules that banks, health providers, and others answer to. Its notification thresholds also mean the smallest foreign services may fall outside the heaviest obligations until their Vietnamese traffic grows. The takeaway is scope discipline: confirm whether you are a covered provider before assuming every clause binds you.

The cross-border read for US firms

For a US operator, the identity rule is the headline. Real-name verification of every active account is a sharp contrast with the US posture, where Section 230 shields platforms and imposes no identity-verification duty, and it goes further than the EU's Digital Services Act, which centers on notice-and-action and risk assessment rather than mandatory user identification. A US platform that lets Vietnamese users sign up anonymously is, on the decree's face, out of step with a binding local rule. The practical exposure is not theoretical: the authentication deadline has already passed.

Who owns the rule now

There is a live uncertainty worth flagging. Decree 147 assigns oversight to the Authority of Broadcasting and Electronic Information under what was the Ministry of Information and Communications. Vietnam began a government restructuring shortly after the decree issued, and Tilleke notes that references to the MIC and that authority are subject to change. Reporting indicates the MIC's functions were folded into the Ministry of Science and Technology in the 2025 merger. Firms should confirm the current supervising body before filing a notification or responding to a removal order, because the named agency in the 2024 text may no longer be the one on the other end of the request.

Frequently Asked Questions

What did Vietnam's Decree 147/2024 change?

It replaced Decree 72/2013 as the master rule for internet services and online information, and it made two duties binding: verifying every active user account against a Vietnamese phone number or national ID, and removing unlawful content quickly, as fast as 24 hours on an authority's request.

Who has to comply?

Domestic and offshore social networks, content platforms, and app providers that either store data onshore in Vietnam or attract more than 100,000 monthly Vietnamese visits over six consecutive months. Covered offshore providers must also notify the authorities within 60 days of crossing that threshold.

Does Decree 147 require labeling AI-generated content?

The decree's confirmed core duties are identity verification and rapid takedown. The explicit mandate to label AI-generated audio, image, and video sits in Vietnam's AI Law (No. 134/2025/QH15) and its 2026 implementing decree, not in Decree 147. Treat 147 as the identity-and-moderation groundwork that came first.

When did compliance start?

The decree took effect December 25, 2024. Providers had 90 days, until March 25, 2025, to authenticate their existing active users, so the core obligations are already in force and past their first deadline.

Browse the full AI Regulation News tracker

Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel.