Your data, honestly.
We're researching data sovereignty while using platforms we don't control. Here's exactly how that works.
Two systems handle your data
What we control (the research)
Your registration details are stored in Supabase (Sydney, AU) with row-level security. Transcripts stay in ElevenLabs — we download them from there for analysis. You can ask us to delete your registration data anytime.
What we don't control (the tech)
ElevenLabs (US) processes your voice. Your audio goes to US servers. We've opted out of model training, but their Terms of Service grant a broad license. We can't revoke that after the fact.
Where your data goes
| Data | Where |
|---|---|
| Registration | Supabase (Sydney) |
| Voice & transcript | ElevenLabs (US) |
| Transcripts (downloaded for analysis) | ElevenLabs (US) |
| Wānanga | Zoom (US) |
Your rights
- Stop anytime — close the browser
- Switch between AI and person whenever
- Withdraw from our database up to two weeks after wānanga
- We delete our copy within 3 years (or earlier on request)
- ElevenLabs retains up to 3 years — we can't control that
- Must be 18+ for the AI; under 18 can book a person
- Not participating has zero consequences
Why we're being this direct
The tools we're using don't align with the data sovereignty principles our research is built on. We chose them because building our own voice AI was beyond our capacity. That contradiction is real — and it's one of the things this research explores.