Is my data used to train AI?

Never. Not audio, not transcripts, not notes. Your consultations are not a dataset.

Not by Akoua, and not by its providers

Patients’ clinical data is never used to train models — not by Akoua, and not by our providers. That holds on every sovereignty profile.

• AU Sovereign. Everything stays in Australia: transcription and note drafting run on vetted Australian services over private links inside Akoua’s cloud. No provider uses your data to train models — model-improvement programmes are opted out on every request, and our providers do not train on customer content. Speaker identity is matched only on Akoua’s own infrastructure and never sent to a provider.
• Global. Audio and transcripts are processed by vetted global providers under your organisation’s consent, and none of them uses your data to train their models.

On both of those profiles your clinical record is stored in Akoua’s Australian cloud in Sydney; on the UK Sovereign region it is stored in the United Kingdom.

What we do use

De-identified operational metrics — counts, performance, reliability — are used to run and improve the service. Nothing clinical. Akoua doesn’t sell personal information, use it for advertising, or profile you, and there are no tracking or advertising cookies and no third-party analytics or advertising SDKs in the apps.

If that ever changed

It would be opt-in, disclosed on the privacy by design page first — and there are no plans to change it.

This is also why the model never has to be clever about your consult: Akoua is a documentation aid, not a diagnostic device. It makes no clinical inferences, drafts from only what was said, marks anything absent as [not stated], and requires your explicit confirmation before a note is saved. The reasoning is set out in the responsible-use statement and the privacy notice for your region.

Related: what happens to the audio from a consult.