Consent and Rights When Cloning a Voice
Voice cloning is a powerful tool. It lets you preserve a speaker's emotional delivery, tone, and personality across translations and new recordings. But that power comes with a responsibility: you need explicit permission from the person whose voice you are cloning.
The legal and ethical lines are not always obvious. What about guest speakers? Interviews? Archive recordings of people no longer available to ask? This guide walks you through who must consent, how to get it in writing, and what to do when the situation is messy.
Who Must Consent
The person whose voice you are cloning must give permission. This applies whether you own the recording or not.
If you recorded someone speaking on your own video, interview, or podcast, you are making a derivative work of their voice. Even if they agreed to be recorded for the original content, cloning their voice for new sentences they never spoke is a new use. You need their explicit consent for that new use.
Here is a concrete example. You film a customer testimonial where Sarah describes how your product solved her problem. The testimonial is yours to use in marketing because Sarah signed a release. But cloning her voice to generate testimonials in Spanish, French, or German without asking is a separate use. Her voice now says things she never recorded. That is a new work that requires new permission.
If you recorded yourself, you can clone your own voice without asking anyone else. But if someone else owns the copyright to the original recording, a studio, a previous employer, a collaborator, you may still need their permission to use it as a reference.
The simplest rule: if you had to get permission to use their face, voice, or name in the original content, you need permission to clone it.
Copyright vs. Personality Rights
These are different. Copyright ownership (who owns the recording file) and personality rights (who owns their voice and likeness) do not always belong to the same person.
You might own the copyright to an interview recording because you paid for the studio and editing. But the guest retains their personality rights. You own the file; they own their voice. Cloning their voice requires their consent, not just your copyright ownership.
Conversely, you might not own the copyright to archive footage of a historical speech, but that person may have granted limited rights to their voice in the original agreement. When you add a new use like voice cloning, you are still stepping over their personality rights, even if you licensed the file itself.
Guests and Interview Subjects
Guest speakers are the trickiest case. When someone appears on your show or gives you an interview, the expectations shift.
If a guest signed a standard talent release, that release usually covers the original recording. It does not automatically cover voice cloning. A release that says "you may use this interview in broadcasts" does not mean "you may clone this person's voice for other purposes." Those are two different things.
Before you clone a guest's voice, go back and ask. Make the ask specific: "We want to use AI voice cloning to dub your interview into five languages. Your voice will synthesize new sentences in those languages, with the same tone and emotion. Do you agree?" Give them a chance to say no.
When a Guest is Hard to Reach
If the guest is no longer reachable, they moved, changed contact info, or the interview was years ago, you cannot assume consent. Here is what to do:
-
Try to reach them. Use email, LinkedIn, social media, or ask mutual contacts. Document your attempt.
-
If they do not respond within a reasonable time (two weeks is standard), you have exhausted good-faith effort.
-
At that point, you have three options: use subtitles instead of dubbing, hire a voice actor who sounds similar but does not aim to clone the original, or do not dub the content at all.
Your safest move when you cannot confirm consent is to use subtitles or a different voice. The legal risk of proceeding is not worth the convenience of avoiding a conversation you could not have anyway.
Special Case: Public Figures and Archive Content
Public figures have weaker privacy protections than private individuals in many jurisdictions, but that does not erase their rights. A news interview of a politician or celebrity still belongs to that person. The fact that the content is public does not grant you permission to clone their voice for new purposes.
If you are working with archive material, historical speeches, or content from decades ago, the original speaker may be deceased or completely unreachable. In those cases, the legal path is to get permission from their estate if one exists, or to accept that cloning is not available and choose an alternative approach.
Writing and Storing Consent
Get it in writing. Email, Slack, a signed form, or a record in your notes all count. A quick email exchange is enough, but you need three clear components:
-
Clear identification. Name the content (the interview, the video, the podcast episode). Include the date or a link so there is no confusion later. "The video we shot on March 15, 2024" is better than "the recent interview."
-
Specific permission for voice cloning. "I agree to have my voice cloned using AI for dubbing this content into other languages" is clear. "I agree to participate" is not enough, and neither is "you can use this content however you want."
-
Scope of use. Be honest about where the cloned voice will appear. "YouTube, Instagram, and the course platform we sell through" is complete. "Anywhere we want" is too vague and feels evasive, and the person may decline or ask for limits.
You do not need a lawyer to write this. A short email works:
"Hi Alex, we are dubbing the interview we recorded on March 15 into Spanish, German, and French using AI voice cloning. Your voice will be preserved in all versions, with the same emotion and tone. The dubbed versions will be published on our YouTube channel and our website. Are you okay with this?"
If they say yes, save that email. Reply asking them to confirm in writing again: "Thanks Alex, just to confirm: yes, you agree that we can clone your voice and publish the dubbed versions on YouTube and our website?" Store both messages together.
You are done once you have that written agreement. Keep the record for as long as the content exists. If the person later objects or legal questions arise, the email proves consent.
Consent for Different Uses
Consent is specific. If you get permission to dub into Spanish only, but later want to dub into Mandarin, you need to ask again. If you get permission to publish on YouTube but later want to license it to a course platform or sell it as part of a product, ask for clarification.
Updating the consent is a quick message: "Since we recorded that email, we have had requests to publish the dubbed version on course platform X as well. Are you okay with that?" Save the response.
When You Cannot Get Consent
Sometimes you cannot get consent. The speaker has passed away. A recording is decades old. You lost contact after multiple attempts.
In these cases, you have several paths:
Use Subtitles
Your audience reads the translation instead of hearing a dubbed voice. This sidesteps the consent issue entirely and is perfectly valid for many content types. Documentaries, educational videos, and archival material often work best with subtitles. The viewer hears the original voice and reads the translation, which is often more authentic than a cloned reconstruction.
Subtitles also work well when the original speaker had a distinctive accent or unique vocal qualities that matter to the content. You preserve all of that without any consent questions.
Use a Different Voice
Dub with a generic AI voice that does not aim to recreate the original speaker. This is a different product, but it is legal and honest. The audience hears that this is a dubbed version because the voice is clearly not the original. Many content creators choose this route for international versions.
The downside is that you lose the personality match. A CEO's natural, authoritative tone becomes a generic read. But this is a trade-off worth making when consent is not available.
Narrator Track
Add a separate narrator voice that delivers the translation. This is common in documentaries and educational content. The original audio plays during visuals, and a narrator reads the translated script over the original. Audiences understand this format and do not expect the original speaker to be cloned.
Do Not Dub
Keep the content in the original language or combine subtitles with the original audio. Some content travels best this way, especially if the speaker's voice is a key part of the appeal.
Never try to clone a voice without permission hoping no one notices. The person, their family, or their estate may object. Platforms may flag it. The trust cost of being caught outweighs the convenience of not asking first. It is also unethical to use someone's voice without permission, regardless of legal risk.
Platform Rules and Disclosure
Most platforms where you publish (YouTube, TikTok, Instagram) are adding rules about synthetic audio and voice cloning. Some require you to disclose that AI was used. Some are stricter.
YouTube
YouTube requires disclosure if you publish content about real people that is either inauthentic, altered, or synthetic. If you clone someone's voice, you must note that clearly in your description. Failure to disclose can result in demonetization or removal.
TikTok
TikTok labels synthetic voices automatically in some regions. If you use a cloned voice, the platform may add a label. This is separate from getting consent, but it shows that platforms are increasingly tracking and flagging synthetic audio.
EU and Other Regions
European Union regulations and other jurisdictions may require explicit labels on all synthetic media, not just voice. Some countries require consent from the original speaker before any synthetic voice can be published, regardless of platform.
Even if you have clear consent from the original speaker, check your platform's policy and your local laws. Getting consent from the speaker does not exempt you from platform rules or legal requirements in your country.
When you do have consent, disclose honestly. "This interview is dubbed using AI voice cloning" or "The international versions use cloned voices" tells your audience what they are listening to. Most audiences are fine with that if you are transparent.
Common Mistakes
Mistake 1: Assuming an Existing Release Covers Voice Cloning
A talent release or permission form that covers the original recording does not automatically cover cloning. Many creators assume it does and get into legal trouble later. Always ask specifically about voice cloning, even if the person signed a form years ago.
Fix: Go back and get explicit written permission for voice cloning specifically, separate from the original recording consent.
Mistake 2: Cloning Without Disclosing to Your Audience
You have consent from the speaker but do not mention that the dubbed version uses AI voice cloning. Your audience feels deceived when they find out.
Fix: Always disclose in your description or title that voice cloning was used. Be transparent about which languages or sections use cloned voices.
Mistake 3: Forgetting to Document Consent
The person verbally agreed, but you did not save the email or written confirmation. Months later, they change their mind or deny they ever agreed.
Fix: Always get written consent and store it. A quick email is enough. Save it for as long as the content exists.
Mistake 4: Expanding the Scope Without Re-Asking
You had permission to dub into Spanish, but now you want to add French or publish on a different platform. You assume the original permission covers it.
Fix: Ask again in writing when the scope changes. The conversation takes two minutes.
Mistake 5: Assuming Archive or Public Content is Fair Game
The content is old or the person is famous, so you assume you can clone their voice.
Fix: Get permission anyway. Fame does not erase rights. Archive age does not erase rights.
Decision Framework: Should You Clone This Voice?
Before you clone any voice, run through this decision tree:
-
Do you have permission from the person whose voice it is? If no, skip to section "When You Cannot Get Consent." If yes, continue.
-
Is the permission in writing? If no, ask for it now via email. If yes, continue.
-
Does the written permission cover the specific language or platform you want to use? If no, ask for clarification. If yes, continue.
-
Have you checked your platform's rules and your local legal requirements? If no, do that now. If yes, you are clear to proceed.
-
Will you disclose to your audience that voice cloning was used? If no, decide now. If yes, you are ready.
If you get stuck at any step, reread the relevant section above or choose an alternative method (subtitles, different voice, narrator track).
FAQ
Q: Can I clone someone's voice if they are dead? A: No, unless their estate has given permission or there is a strong legal reason to do so (like restoring historical content). Contact the estate if one exists. Otherwise, use subtitles or a different voice.
Q: If I own the copyright to a recording, do I automatically own the right to clone the voice? A: No. Copyright ownership and personality rights are different. You might own the file but not own the right to clone the voice. Always ask the original speaker.
Q: Does a verbal agreement count as consent? A: It is risky. Always get written consent via email or a signed form so you have proof later.
Q: What if the person says they were not comfortable with voice cloning after I already published? A: Take the video down. Apologize. Offer to re-publish with subtitles or a different voice instead. Do not argue about what they originally agreed to.
Q: Do I need to tell a voice cloning platform like DubLab that I have consent before I upload? A: Most platforms have their own terms of service. Read them. Some platforms may require you to certify that you have necessary rights. Always follow the platform's guidance on consent and disclosure.
What to Do Next
Here is your action plan before you clone any voice:
-
Identify all people whose voices appear in your content. Make a list of every speaker, guest, or person who will have their voice cloned.
-
Decide for each person: do I have clear, written permission for voice cloning specifically? If yes, move to step 3. If no, choose an alternative method (subtitles, different voice, narrator track) for that person or segment.
-
Get written permission from anyone you do not have it from yet. Send them a specific, honest ask via email. Save their response.
-
Check your platform's rules and your local legal requirements. Make sure voice cloning is allowed and that you know what disclosures are required.
-
Create your dubbed version. Clone the voices you have permission for, and use alternatives for anyone you do not.
-
Disclose clearly. In your description, video end-card, or on screen, note that voice cloning was used. Name the languages or segments that were cloned.
-
Keep your consent records for as long as the content exists. You may need them if a question arises later.
The consent conversation takes minutes. Sorting out problems later takes days and erodes trust with your audience and with the people whose voices you are using.
🚀 Start Dubbing Your Videos Today
DubLab uses AI to translate your videos into 92+ languages in minutes.