Responsible AI
Voice cloning is powerful. Used carelessly, it's a deepfake factory. AudioPod is built around the opposite premise — explicit consent for every clone, real human review for the edge cases, and a transparent channel for misuse.
Our position
Voice is the most personal piece of identity most people will ever digitize. We treat it that way. Every voice clone on AudioPod has a creator who confirmed they own the voice or hold a written license, and a record of that confirmation that we can produce on a legitimate request.
We won't clone political figures, public officials, or non-consenting public figures — full stop. We won't sell synthetic-voice access to obvious impersonation-for-fraud use cases. And we won't use customer audio to train our models unless customers opt in with their eyes open.
Safeguards
Not all of these are unique to AudioPod. All of them are necessary. We document them publicly so creators, partners, and regulators can hold us to them.
Upload
Voice sample
Verify
Identity check
Sign
Consent record
Lock
Owner-only use
Consent verification
Before a voice clone can be used, the creator confirms they own the voice or hold a written license. Every cloned voice has a consent trail.
No public-figure cloning
We block clones of political figures, public officials, and non-consenting public figures. Flagged uploads route to human review.
Provenance signals
AudioPod-generated audio carries provenance metadata so downstream platforms and detectors can identify synthetic content.
Takedown SLA
Misuse and impersonation reports are triaged within 24 hours. Verified takedowns complete within 72 hours.
Audit trail
Every cloned-voice generation is logged with timestamp, prompt, and account context. Available on legitimate request.
Human review
Flagged generations and consent disputes are escalated to a human reviewer — not silently auto-approved by a classifier.
What we won't do
Report misuse
We triage trust reports within 24 hours and complete verified takedowns within 72 hours. No bots, no auto-responders — a person reads every report.
Email the trust team
Send the suspected output, the original recording (if you have one), and any context that helps. We'll come back within a business day.
trust@audiopod.aiTriage
Within 24 hours
Suspension
During investigation
Takedown
Within 72 hours
Common questions
Adjacent policies
Responsible AI is the ethics layer. Trust covers operations and security. The manifesto says why.
Bookmark this page. Send the link to your legal team. If you find a gap between what we promised and what we shipped, write to the trust team — that's the most useful email we get.