Deepfake laws: US, EU and what platforms must do
Deepfake rules are arriving in pieces: a US federal law on nonconsensual intimate imagery with platform duties already in force, EU labelling duties and bans coming in stages, and proposals on digital replicas. This page sets out what we could confirm from primary sources. It is a plain-language summary, not legal advice. Section hub: AI law, copyright and safety; the EU rules are explained in the EU AI Act explained.
By the Chelure Enterprises Editorial Team. Last reviewed . How we check our sources.
At a glance
| Jurisdiction | Rule | Key dates and numbers |
|---|---|---|
| United States | TAKE IT DOWN Act: covers nonconsensual intimate real images and 'digital forgeries' created or altered with software, an app or AI; covered platforms must run a notice-and-removal process | Section 3 effective May 19, 2026; removal within 48 hours; civil penalties of $53,088 per violation (FTC) |
| European Union | AI Act Article 50(4): deployers must disclose deepfakes; new Article 5 bans on non-consensual sexual deepfakes and child sexual abuse material | Article 50 from 2 August 2026; new bans from 2 December 2026 |
| United States (proposal) | Copyright Office Part 1 recommended a federal digital replica law | Part 1 published July 31, 2024 |
United States: the TAKE IT DOWN Act
According to the FTC's business guidance, the Act's definition of covered platform covers various websites, apps and online services, with examples including social media, messaging, image or video sharing and gaming platforms. It requires removal, within 48 hours of a valid request, of nonconsensual intimate photos and videos, including digital forgeries, and of known identical copies. Victims do not need to report duplicates, but platforms must make reasonable efforts to find them. Platforms must provide a clear and conspicuous plain-language notice explaining how to submit a request; the FTC also recommends giving each request an identifying number and telling requesters whether content was removed or why not. The FTC enforces the requirements, and violators may face civil penalties of $53,088 per violation. The FTC's guidance says it does not cover every requirement, so read the Act itself for the full text.
European Union
Article 50(4) of the AI Act requires deployers to disclose that deepfakes are artificially generated or manipulated; for evidently artistic, creative, satirical or fictional works the duty is limited to an appropriate notice that does not hamper enjoyment of the work. Separately, the 2026 Digital Omnibus added Article 5 prohibitions, applying from 2 December 2026, on AI systems that generate or manipulate realistic images, video or audio of an identifiable person's intimate parts or sexually explicit activities without consent, and on systems that generate child sexual abuse material. Details and dates: explained and timeline.
Replicas of voices and faces
The Copyright Office's Part 1 report addressed digital replicas and recommended a federal law; see AI and copyright law. We have not verified other countries' deepfake statutes on primary sources, so we do not list them.
If you run a platform or publish synthetic media
- Platforms: publish a clear takedown process and be able to remove reported content and known copies within 48 hours.
- Publishers: label synthetic media; consider provenance data such as C2PA credentials, noting they can be stripped.
- Developers: do not build or offer tools that produce non-consensual sexual imagery; EU bans start on 2 December 2026.
Frequently asked questions
Is there a US federal law on deepfakes?
The TAKE IT DOWN Act covers nonconsensual intimate imagery including 'digital forgeries' made or altered with software, an app or AI. Its notice-and-removal requirements for covered platforms took effect on May 19, 2026, and the FTC enforces them.
What must platforms do under the TAKE IT DOWN Act?
Remove reported content, and known identical copies, within 48 hours of a valid removal request, and give clear and conspicuous plain-language information about how to submit a request, according to the FTC's business guidance.
What are the penalties?
The FTC says violators may face civil penalties of $53,088 per violation, because a violation is treated as a violation of an FTC rule.
What does the EU require?
Article 50(4) of the AI Act requires deployers to disclose deepfakes as artificially generated or manipulated, with a lighter notice for evidently artistic, satirical or fictional works. From 2 December 2026, new bans in Article 5 cover AI systems that generate sexually explicit realistic material of an identifiable person without consent, and child sexual abuse material.
What about digital replicas of voices and faces?
The US Copyright Office's Part 1 report (July 31, 2024) addressed digital replicas and recommended a federal digital replica law. We have not summarised other countries' deepfake laws here.
Sources
- FTC: Complying With the Take It Down Act
- AI Act Explorer: Article 50, transparency obligations
- AI Act Explorer: Article 5, prohibited practices
- US Copyright Office: Artificial Intelligence Study
Primary sources opened on November 13, 2026. Prices come from our verified price table (see the methodology).