Fact · Dictation software
Detta's privacy policy says it uses uploaded Shared Content, which includes recordings, to train its speech models.
Quote found word for word on the live source, last checked October 23, 2026
In the source's own words
“We use Shared Content to train and evaluate our speech recognition models and to improve Detta”
Detta Privacy Policy, Fermion Research, Fermion Research, 2026, Section 6.1, Shared Content.
Open the original- What it measures
- Shared Content means the Recordings and Dictation Records that Detta uploads under the Sharing Setting (Section 1(i)).
- What it does not show
- Only when the Sharing Setting is on, which Section 7.3 says is on by default. Turning it off in Settings stops further uploads.
In the data
A row of Dictation App Features Compared: Offline, Privacy, AI, as Detta, recordings train models: 1 1 = yes, 0 = no.
- App
- Detta
- Feature
- Recordings can train the maker's models
- Detail
- opt-out in Settings, on by default
- As of
- 2026-10
- Note
- 1 means the maker's page says the app does this. The quoted sentence says uploaded Shared Content, which includes recordings, trains the models, and the setting can be turned off.