Quick Answer
Copyright is the wrong tool here: a voice as such is not protected the way a recording is, which is why the U.S. Copyright Office recommended a new federal digital replica law in its report of 31 July 2024. What protects a voice today is likeness law. Tennessee’s ELVIS Act, in force since 1 July 2024, defines “voice” to include a simulation of it, while the federal NO FAKES Act of 2026 was only reported to the Senate on 24 June 2026 and is not law. In the EU, the AI Act applies from 2 August 2026 and requires deepfake content to be disclosed. Clone your own voice, or hold written consent, and disclose realistic output.
Sources read 16 September 2026: the EU AI Act applies from 2 August 2026, and the NO FAKES Act of 2026 remains a pending bill. Information, not legal advice.
Voice cloning rules with sources and dates: why copyright is the wrong claim, what the ELVIS Act covers, why the NO FAKES Act is still only a bill, and what the EU AI Act requires from 2 August 2026.
Free tool, runs in your browser
Check your voice sample first, free
A clone is only as good as the sample. The checker rates length, noise, clipping and silence in your browser, nothing uploaded. Cloning itself is free; speaking with the voice costs 2 credits for a clip up to 40 words.
Open the tool →
Related next steps
This page collects what the rules actually say about cloning a voice, with a source and a date for each one, as read on 16 September 2026. It is information rather than legal advice, and none of it is a substitute for a lawyer who knows your jurisdiction.
Copyright is the wrong question
People search for "AI voice copyright", but copyright protects fixed works such as a particular recording, not the sound of a person. That gap is not a matter of opinion: the U.S. Copyright Office published Part 1 of its AI report ↗ on 31 July 2024 specifically on digital replicas, and recommended a federal digital replica law. A recommendation to create a new right is evidence that the existing one does not cover this. Part 2, on copyrightability, followed on 29 January 2025.
So if someone clones your voice, the claim you are looking for is usually not copyright infringement. It is a likeness or publicity claim.
What protects a voice today, in the United States
Right of publicity law is made state by state, and one state wrote the AI case into statute first. Tennessee's ELVIS Act, the Ensuring Likeness Voice and Image Security Act, was signed on 21 March 2024 and took effect on 1 July 2024 ↗. Three parts of it matter to anyone using these tools:
- It defines "voice" to include an individual's "actual voice" and a "simulation of the individual's voice", so a convincing imitation is inside the statute.
- It prohibits publishing, performing or distributing a voice or likeness "with knowledge that use of the voice or likeness was not authorized".
- It also bars distributing "an algorithm, software, tool or other technology" whose purpose is producing a particular identifiable individual's voice or likeness without authorization. That clause reaches tool makers, not only posters.
Violations carry civil liability and criminal exposure as a Class A misdemeanor, up to 11 months and 29 days and fines to $2,500.
The federal bill, and its actual status
The NO FAKES Act of 2026, S.4591, whose official title is the Nurture Originals, Foster Art, and Keep Entertainment Safe Act of 2026 ↗, would create a federal intellectual property right against unauthorized digital replicas of a person's voice and visual likeness, with a notice and takedown process, a counter notification process for contesting removals, and exemptions covering news reporting, sports broadcasts and satire.
It was reported to the Senate on 24 June 2026. It is not law. Any page telling you that a federal digital replica right already governs your work is wrong as of 16 September 2026, and any tool claiming compliance with it is claiming compliance with a bill.
In the EU, disclosure is already required
The EU AI Act entered into force on 1 August 2024 and applies from 2 August 2026 ↗, which is to say it applies now. Article 50 ↗ requires deployers of an AI system that generates or manipulates image, audio or video content constituting a deep fake, meaning content resembling real people that "would falsely appear to a person to be authentic or truthful", to "disclose that the content has been artificially generated or manipulated".
Providers whose systems generating synthetic audio, image, video or text were already on the market before 2 August 2026 had to meet the related provider duty by that date. For a creator, the practical effect is simple: in the EU, a realistic cloned voice in published content is disclosed, not quietly shipped.
The consumer protection angle
Separately from likeness law, the FTC proposed a rule in February 2024 ↗ that would prohibit the impersonation of individuals, extending its rule on government and business impersonation. It also sought comment on whether it should be unlawful for a firm, such as an AI platform, to provide goods or services it "knows or has reason to know" are being used to harm consumers through impersonation. Voice cloning was named explicitly in the announcement.
What to do, in practice
- Clone your own voice, or hold written consent. On Percify, Clone Yourself makes your voice from a sample of 15 to 30 seconds for 5 credits, once. If the voice belongs to someone else, get consent that names AI voice cloning and the uses you intend, and keep it.
- Do not clone a named public figure for commercial content. That is the exact fact pattern the ELVIS Act and the NO FAKES bill were written for.
- Disclose realistic output. The EU requires it under Article 50, and TikTok, YouTube and Meta each require their own label, collected in each platform's label rule.
- Keep the sample and the consent together, so the provenance of a voice can be shown later.
For the tools themselves, including which ones you can run yourself, see cloning your voice free; Percify's prices are on plans and packs.
None of this is legal advice, and the law here is moving: the federal bill is pending, and state statutes continue to be written.
Ready to Create Your Own AI Avatar?
Make a talking video of yourself from one photo, in your own voice. No watermark on any plan.
Cancel anytime from your profile
Got questions?
Frequently asked
Not the voice itself. Copyright protects fixed works, such as a specific recording, rather than the sound of a person. The U.S. Copyright Office published a report on digital replicas on 31 July 2024 recommending a new federal law for exactly this gap, which is why voice claims usually rest on likeness and publicity law instead.
It depends on whose voice it is and where you are. Cloning your own voice is the straightforward case. Cloning someone else's without authorization runs into state likeness law, and Tennessee's ELVIS Act, in force since 1 July 2024, defines "voice" to include a simulation and carries civil and criminal exposure. In the EU, realistic AI content must be disclosed under the AI Act, which applies from 2 August 2026. Read on 16 September 2026; this is information, not legal advice.
Not yet. The NO FAKES Act of 2026 (S.4591) would create a federal right against unauthorized digital replicas of voice and likeness, with a notice and takedown process, but it was reported to the Senate on 24 June 2026 and has not been enacted as of 16 September 2026.
Treat it as required. Tennessee's statute turns on use "with knowledge that use of the voice or likeness was not authorized", and the pending federal bill is built around authorization as well. Get consent that names AI voice cloning and the intended uses, in writing, and keep it with the voice sample.
For realistic content, yes. EU AI Act Article 50 requires deployers to disclose that deepfake content, meaning generated or manipulated audio or video resembling real people that would falsely appear authentic, has been artificially generated or manipulated. The Act applies from 2 August 2026.
