Shadow 06 / Figure 7
Who must mark what
Scroll the table sideways to see every column.
| Jurisdiction | Instrument | Issuer | Covered actor | Medium | Duty | Status | Relevant date (type stated) | Source |
|---|---|---|---|---|---|---|---|---|
| 1 Statute / regulation | ||||||||
| China | Labelling Measures and later CAC programmes | Cyberspace Administration of China | — | — | — | official text located (Chinese) located, not cleared | — | cac.gov.cn |
| European Union | AI Act Art. 50 (as amended by Reg. 2026/1744) | EU | providers; deployers | generative output; deepfakes; public-interest text | providers: machine-readable marking (exceptions incl. assistive standard editing). Deployers: disclose deepfakes (adapted for evidently artistic, creative or satirical work); disclose AI-generated text published to inform the public on matters of public interest, unless it has undergone human review or editorial control AND a person holds editorial responsibility. That exception applies only to the text disclosure, not to deepfakes or provider marking | applies | applicable from 2 Aug 2026; legacy systems by 2 Dec 2026 | EUR-Lex; AI Act Service Desk |
| India | IT Rules amendment 2026 | MeitY | intermediaries; SGI tools | synthetically generated information (text-only excluded) | label and provenance for SGI | in force | in force 20 Feb 2026 | MeitY FAQ |
| Kazakhstan | AI Law No. 230-VIII | Parliament | — | — | — | record located; full text not retrieved located, not cleared | — | adilet.zan.kz |
| Russia | Federal Law 243-FZ | Federal Assembly | — | — | — | record located; operative text not verified located, not cleared | — | publication.pravo.gov.ru |
| South Korea | AI Basic Act + MSIT guidance | National Assembly; MSIT | providers | outputs; realistic synthetic content | marking and notice, differentiated by output type | commenced; grace period announced | commenced 22 Jan 2026 | MSIT |
| United States — California | AI Transparency Act as amended (SB 1000; AB 853; AB 2713) | State of California | providers (no user threshold); covered platforms from 2027 | image, video, audio (latent disclosure) | latent disclosure; free disclosure verification tool | operative | operative 2 Aug 2026; platforms from 1 Jan 2027; capture devices from 2028 | California legislature (chaptered texts) |
| Vietnam | Law on AI No. 134/2025/QH15 | National Assembly | providers; deployers | specified generated or edited outputs | providers: machine-readable marking; deployers: disclosure (Art. 11 conditions) | in force | in force 1 Mar 2026 | Government of Vietnam (English text) |
| Brazil | PL 2338/2023 (general AI bill) | Chamber of Deputies | — | — | — | pending in Chamber located, not cleared | — | camara.leg.br |
| 2 Election rule | ||||||||
| Brazil | TSE Resolution 23.755/2026 | Superior Electoral Court | — | — | — | official text located (Portuguese) located, not cleared | — | tse.jus.br |
| Japan | Revised Public Offices Election Act (Law No. 58) | Diet | — | — | — | official record located (Japanese) located, not cleared | — | sangiin.go.jp |
| Singapore | Prohibition of realistic but false candidate depictions | Elections Department | publishers incl. redistribution | audio, visual, audiovisual | prohibited during election period (a label does not exempt) | in force | applies during the election period | Elections Department Singapore |
| 3 Sectoral / media standard | ||||||||
| Australia | eSafety DIS Standard | eSafety Commissioner | high-impact generative-AI services | AI outputs | differentiate AI outputs | in force | — | eSafety Commissioner |
| Colombia | Territorios al Aire 2026 programme conditions | MinTIC | — | — | — | official text located (Spanish) located, not cleared | — | mintic.gov.co |
| UAE | Media content standards (Council statement, Sep 2025) | UAE Media Council (2025); National Media Authority, its successor (restatement, Sep 2026, as reported) | media and content creators | AI depictions of national symbols and public figures | prior official approval for AI depictions of national symbols and public figures; AI misuse to spread misinformation, incite hatred or defame others subject to the media violations regulation | statement reported by WAM; restatement reported by press | statement 25 Sep 2025; restatement late Sep 2026 | WAM; Khaleej Times; Gulf News |
| 4 Guidance / policy | ||||||||
| Australia | Being clear about AI-generated content | National AI Centre | businesses | — | discusses labelling, watermarking, metadata | voluntary | published Nov 2025 | industry.gov.au |
| Bahrain | General Policy for the Use of AI | Information & eGovernment Authority | — | — | general AI policy; marking clauses not verified | issued cleared (issuance) | issued Jul 2025 | iga.gov.bh |
| Egypt | National Guidelines for Generative AI | Government of Egypt (ai.gov.eg) | — | — | — | located; full text not retrieved located, not cleared | published Mar 2026 | ai.gov.eg |
| Indonesia | Public recommendation by a ministry official | Communications ministry | content creators | AI-generated content | label or disclaimer (recommended) | recommendation | statement Jun 2025 | ANTARA |
| Kenya | Media guide on the use of AI | Media Council of Kenya | media enterprises | — | inform users of AI use | sectoral guidance | — | Media Council of Kenya |
| Malaysia | National Guidelines on AI Governance & Ethics | MOSTI | — | — | disclosure among suggested measures | guidance | — | MOSTI |
| Mexico | Recommendations for election coverage | INE | — | — | — | official text located (Spanish); non-binding located, not cleared | issued Aug 2026 | ine.mx |
| Oman | General Policy for the Safe and Ethical Use of AI Systems | MTCIT | — | — | general AI policy; marking clauses not verified | published cleared (publication) | published 1 Apr 2025 | mtcit.gov.om |
| Qatar | AI ethical guidelines | MCIT | — | — | — | voluntary, non-binding (self-stated) | — | MCIT |
| Saudi Arabia | Generative AI public guidance; deepfake guidance | SDAIA | — | — | — | located; full text not retrieved located, not cleared | — | sdaia.gov.sa |
| Ukraine | Guidelines for responsible AI use in media | National Council (with ministries) | media | — | clear separation of human-made and AI-generated content (recommended) | guidance | — | National Council of Ukraine |
Instruments we located and verified. A comparison of specified obligations, not a ranking of countries' overall AI regulation. Grey rows: official texts located but not cleared for detailed description; only metadata is shown.
Note
United Kingdom: a January 2026 House of Commons Library briefing stated that no UK legislation requires AI-generated content to be labelled.
SOURCE · Official instruments and issuer publications named in the Source column.