Consent & ethics
A voice is sacred. We treat it that way.
Aeterna can recreate a storyteller's voice so families can hear chapters in their own words. That power deserves clear rules.
Why we treat voice differently
A voice carries identity in a way text never will. It can comfort, persuade, and — in the wrong hands — deceive. That's why every voice clone in Aeterna is built on explicit, revocable consent and protected by hard technical limits.
Explicit consent, always
A voice clone is only created after the storyteller has read and agreed to a plain-language consent form. They can revoke consent at any moment, and the voice model is then deleted.
How consent works, step by step
1. Plain-language consent form
Before any voice training, the storyteller reads a one-page form in their own language explaining what a voice clone is, what it can be used for, and how to revoke it.
2. Active confirmation
They confirm with a typed name and a recorded sentence in their own voice. No pre-checked boxes, no buried links.
3. Limited initial scope
By default, the voice is only used to read back the storyteller's own chapters to invited family. Conversational AI ("Talk to them") is a separate, opt-in setting.
4. Revocation, anytime
The storyteller (or, with proof, their legal representative) can revoke consent at any moment. The voice model is deleted within 24 hours and never used again.
What voice clones can — and can't — do in Aeterna
Allowed
- · Read back chapters of the storyteller's own memoir to invited family.
- · Power "Talk to them" conversations grounded in the memoir, when the storyteller has opted in.
- · Animate a single approved photo speaking a memoir excerpt, with the storyteller's permission.
Never allowed
- · Generating speech the storyteller never agreed to share.
- · Impersonation, advertising, political messaging, or scams.
- · Licensing voices to third parties or any external use.
- · Creating voices of public figures or anyone without verified consent.
Owner-controlled
- The storyteller decides who can listen and who can converse.
- Voice playback can be disabled per chapter or globally.
- Family access can be revoked at any time.
What we won't do
- We won't generate speech the storyteller never agreed to share.
- We won't use voices to impersonate, advertise, or deceive.
- We won't license voices to third parties.
What happens after the storyteller is gone
This is the question families ask most. Our default is simple: the voice continues to exist only as the storyteller agreed. Before death, they can choose whether the voice and "Talk to them" mode remain available to family, are limited to playback of recorded chapters only, or are fully retired. Family cannot override the storyteller's wishes — even with grief, even with permission from other relatives.
If you're a family member
If you're worried about how a voice clone is being used, or you believe a clone has been created without proper consent, please write to us immediately. We'll investigate, freeze the voice if needed, and respond within two working days.
How we review ourselves
We meet quarterly as a team to review every voice-related feature against these principles. If we can't honestly say a feature respects the storyteller's voice and consent, we don't ship it — even if it's technically possible.
Reporting concerns
If something feels wrong, write to ethics@aeterna.app. We read every message.