AI media rules by country
As of 4 November 2026. A summary for creators, not legal advice.
How ClairSOLO meets these rules
- A visible AI label on screen for the whole episode, in the episode’s language and in English.
- Opening and end cards: “This drama was made with AI. All characters are fictional.”
- An AI flag in the file and a provenance record with every episode.
- Each platform’s AI disclosure set when you publish.
- No real person’s likeness without their signed release.
- Election and news deepfakes prohibited by our acceptable-use policy.
Thailand Draft
Status. Draft AI Act (ETDA under MDES). The public hearing closed on 14 August 2026; not yet enacted.
What it requires. The draft requires machine-readable marking of AI-generated or modified content, clear disclosure where content could mislead (for example elections, investment, food or drug claims, impersonation) and labelling by platforms, with fines of up to THB 5 million. Already in force: a Consumer Protection Board notification requiring AI disclosure in advertising. The NBTC has issued non-binding AI guidelines for telecom licensees (visible labels and machine-readable metadata) and has a master plan to bring OTT, streaming and user-generated video under regulation. Deepfakes used to defame are already a Criminal Code offence.
How ClairSOLO meets it. On-screen label in Thai and English for the whole episode, opening and end cards, an AI flag in the file and a provenance record, platform disclosure on publish, no real likeness without a release, and no election or news deepfakes.
Sources:
summary by Baker McKenzie ·
summary by Norton Rose Fulbright ·
summary by BDO Thailand ·
summary by Mondaq (NBTC guidelines) ·
summary by Mondaq (advertising) ·
summary by The Star
Singapore In force
Status. In force: the Elections (Integrity of Online Advertising) Act 2024 and the online safety laws. There is no general AI-labelling law.
What it requires. Election deepfakes are prohibited, and online election advertising made with AI must be disclosed.
How ClairSOLO meets it. Election and news deepfakes are prohibited by our acceptable-use policy; every episode carries the on-screen AI label, opening and end cards, an AI flag in the file and a provenance record.
Source: summary by Bower Group Asia
Malaysia In force
Status. In force from 1 June 2026: the MCMC Risk Mitigation Code under the Online Safety Act 2025.
What it requires. Licensed platforms must label AI-generated or altered content, including deepfakes.
How ClairSOLO meets it. Each platform’s AI disclosure is set when you publish, and the label, cards, AI flag in the file and provenance record travel with the episode.
Sources:
summary by Malay Mail ·
summary by New Straits Times
Indonesia Draft
Status. Two Presidential Regulations on AI (a roadmap; ethics and governance) were reported in September 2026 to be awaiting the President’s signature. In force: Komdigi Circular Letter No. 9/2023 on AI Ethics.
What it requires. The minister has said platforms will need to label or watermark AI content.
How ClairSOLO meets it. A visible label and machine-readable AI flag on every episode, a provenance record, and platform disclosure on publish.
Sources:
summary by Kontan ·
summary by Indonesia Business Post
Philippines Bills pending
Status. Bills pending, for example Senate Bill 758 on AI-content disclosure and House Bill 10227 requiring deepfake disclaimers. In force: COMELEC Resolution 11064.
What it requires. COMELEC Resolution 11064 prohibits election deepfakes and requires labels. The bills would require disclosure of AI content and disclaimers on deepfakes.
How ClairSOLO meets it. No election or news deepfakes under our acceptable-use policy; the AI label, opening and end cards, AI flag and provenance record on every episode.
Sources:
summary by Foundation for Media Alternatives ·
bill text, House of Representatives (HB 10227) ·
summary by MLex
Elsewhere
China. Labelling measures requiring explicit and implicit labels, in force since 1 September 2025.
European Union. AI Act Article 50 transparency duties apply from 2 August 2026.
Vietnam. The AI law’s labelling duties apply from 1 March 2026 (secondary source; to be verified).
Sources:
summary by Bird & Bird ·
summary by AI Label (secondary)
Rules change quickly. We review this page monthly. Tell us about a change: contact.