The world is writing deepfake law in real time

How are governments regulating deepfakes?

• Mandatory labeling & watermarking

  The dominant model. The EU AI Act (Art. 50), China's Sept 2025 labeling measures, South Korea's AI Basic Act, and India's synthetic-content rules all converge on visible labels plus embedded, machine-readable provenance — pushing duties upstream to model and tool providers.

• Criminalizing non-consensual intimate deepfakes

  The fastest-moving front, and the frontier is shifting from distribution to creation and even possession — the US TAKE IT DOWN and DEFIANCE Acts, the UK's new creation offense, South Korea's expanded sex-crime law, and Australia's federal and state offenses.

• Election-integrity rules

  Targeted bans and disclosure for deceptive political deepfakes near elections — California, Texas, and Singapore's ELIONA — though several US measures face First Amendment challenges.

• Platform liability & rapid takedown

  Laws increasingly put obligations on intermediaries: 48-hour NCII removal in the US, removal directions with seven-figure fines in Singapore, and user-declaration and detection duties in India and California.

• Likeness as a right + provenance

  Denmark's copyright-of-likeness approach and US digital-replica laws treat a person's face and voice as ownable, while provenance standards like C2PA become the technical backbone regulators reference — driving demand for detection, liveness, and authentication.

Timeline

The law, most recent first.

1. Feb 2026

   United Kingdom

   Data (Use and Access) Act 2025, s.138

   Criminalizes creating — and requesting others to create — non-consensual intimate deepfakes, moving beyond distribution.

2. Jan 2026

   United States

   DEFIANCE Act

   A federal civil cause of action for victims of non-consensual sexually explicit deepfakes to sue creators and distributors.

3. Jan 2026

   South Korea

   AI Basic Act

   A comprehensive national AI framework requiring watermarking/labeling of generative-AI output, plus transparency duties.

4. Oct 2025

   India

   IT Rules — Synthetically Generated Information

   Tool providers must mark synthetic outputs with permanent metadata; visible labels must cover at least 10% of the screen, with platform verification duties.

5. Sep 2025

   China

   AI-Generated Content Labeling Measures (GB 45438-2025)

   Mandates both visible labels and hidden metadata/watermarks on all AI-generated media distributed on Chinese platforms.

6. Jun 2025

   European Union

   Denmark — Copyright Act (digital likeness)

   A first-of-its-kind right over your own face, voice, and body, enabling takedowns, compensation, and platform liability for unauthorized deepfakes.

7. May 2025

   United States

   TAKE IT DOWN Act

   Federal crime to publish non-consensual intimate imagery, real or AI. Platforms must remove reported NCII within 48 hours.

8. May 2025

   Japan

   AI Promotion Act + AI guidelines

   A light-touch, principles-based approach: no dedicated deepfake statute yet, relying on AI business guidelines plus existing defamation, copyright, and portrait-rights law.

9. Mar 2025

   Spain

   Draft AI labeling law

   Would mandate clear labeling of AI-generated content, with fines up to €35M or 7% of turnover, aligning national enforcement with the AI Act.

10. Jan 2025

    Canada

    Bill C-63 (Online Harms Act)

    Would have addressed deepfake and NCII harms but died when Parliament dissolved; Canada currently lacks a dedicated federal deepfake statute.

11. Oct 2024

    Singapore

    Elections (Integrity of Online Advertising) Act — ELIONA

    Bans AI-generated content that falsely depicts election candidates during the election period; platforms face fines up to S$1M.

12. Sep 2024

    United States

    California & Texas (state)

    California's 2024 package targets election deepfakes and digital replicas; Texas's TRAIGA (effective Jan 2026) bans unlawful deepfakes with AG-enforced penalties.

13. Sep 2024

    South Korea

    Sex-crime law (deepfake amendments)

    Criminalizes producing, distributing, possessing, or even viewing sexually explicit deepfakes without consent.

14. Aug 2024

    European Union

    EU AI Act (Reg. 2024/1689), Article 50

    Providers must machine-mark AI-generated content; deployers must visibly label deepfakes and AI text on matters of public interest. Penalties up to €15M or 3% of global turnover.

15. Aug 2024

    Australia

    Criminal Code Amendment (Deepfake Sexual Material) Act 2024

    Federal offense for non-consensual sharing and creation of sexually explicit deepfakes, reinforced by state laws in NSW and SA.

16. May 2024

    France

    Loi SREN (2024)

    Criminalizes sharing non-consensual sexual deepfakes and requires that AI-generated content be labeled as such; complements EU AI Act obligations.

17. Feb 2024

    Brazil

    Superior Electoral Court resolution + AI Bill (PL 2338)

    Bans deepfakes in election campaigns and requires AI-content disclosure; a broader AI framework (PL 2338/2023) is advancing through Congress.

18. Oct 2023

    United Kingdom

    Online Safety Act 2023

    Made it a criminal offense to share or threaten to share intimate images, explicitly including deepfakes.

19. Jun 2023

    Taiwan

    Criminal Code & sexual-privacy amendments

    Criminalizes creating and distributing non-consensual sexually explicit deepfakes, following high-profile deepfake abuse cases.

20. Jan 2023

    China

    Deep Synthesis Provisions

    Requires conspicuous labeling of synthetically generated or altered media and prohibits deceptive deep-synthesis content.

By region

Informational only, not legal advice. Laws change quickly and details vary by jurisdiction — verify with qualified counsel before relying on any point here.

Stay ahead of labeling, provenance, and detection mandates