These guidelines explain, in practice, whose voices you may dub with DubForge, how to show you have permission, and how to tell your audience. They are part of the Terms of Use, together with the Acceptable Use Policy.
Your own voice
If you are the person speaking, you may dub your videos into any language DubForge offers. You still confirm the project declaration for each file. After the first dub you may mark which voice in the project is yours; only that voice may enter our study archive.
Someone else's voice
You need the person's written permission before you upload the file. It can be a signed document, an e-mail or a message, as long as it says clearly:
- who gives the permission and who receives it;
- which videos or recordings it covers, one by one or by kind, such as "every lesson of this course";
- that the voice will be reproduced with AI in other languages;
- where the dubs will be published, and for how long;
- whether and how the person may take the permission back.
A general "you can use my videos" does not cover dubbing a voice with AI.
Keep a copy of the permission for as long as the dubs are published and for 5 years after. We may ask to see it. If you cannot show it, we may remove the project.
If the person takes the permission back, stop publishing new dubs of their voice, and delete their projects as you agreed with them.
Several people in one video
Every voice in the file needs permission. If someone cannot or will not give it, do not upload the file, or cut their part out before you upload it. Voices other than your own never enter our study archive.
Working for clients
If you dub videos for clients, each client must have the right to give you permission for every voice in their videos, and you need that permission in writing. The owner of the DubForge account answers for each declaration, also for files sent through the API.
Voices you may not use
- Public figures, without their written permission. Not even for parody, commentary or content that "is obviously fake".
- People who have died, unless whoever holds the rights to their voice and image gives written permission.
- Children, unless a parent or guardian gave written permission, and never in content that could harm them.
- Voices from films, series, games, music, podcasts and other works you do not hold the rights to. The speaker's permission does not replace the rights to the work.
What the project declaration means
Before the first dub of each file, you confirm one statement:
- "I have the right to use every voice in this file: it is my own, or I have written permission from each person who speaks."
It covers every voice in the file, public figures included: without their written permission, their voice may not be in it. It also means you follow these guidelines, telling your audience included.
We record your answer with the fingerprint of the file, so it holds for that exact file and no other. A new or replaced file needs its own declaration. A false declaration breaks the Terms of Use.
Telling your audience
Every dub you export is marked as made with AI, in the file's provenance metadata (C2PA) and in a watermark in the audio. Tell people too, where they will see it: in the description, on the screen at the start, or in the subtitles. For example: "Dubbed into English with AI from the original in Portuguese."
Do not present a dub as a recording the speaker made in that language. Many platforms have their own label for content made or changed with AI; use it when it applies.
Check the translation
You are the one who publishes the dub. Listen to it and read the translation before you publish, above all for names, numbers and technical terms, and for anything where a mistake could hurt someone.
If someone used your voice
Write to support@dubforge.io with "Report" in the subject, saying where the content is. We answer within 48 hours.