Skip to main content

Data Usage: how AI model training works

Training is opt-in, off by default, and managed by your workspace's administrators.

Written by Simon

Our speech AI gets better with real-world examples. Workspaces that opt in help it handle new accents, languages, and recording conditions. Nothing is used for training unless a workspace chooses to opt in.


Is my workspace's data used to train AI models?

Not unless your workspace opts in. Training is opt-in, and no workspace is opted in by default. The same rules apply on every plan.

A workspace can use HappyScribe in full without opting in. Not opting in doesn't affect the accuracy or availability of HappyScribe.


What happens if a workspace opts in

Content from an opted-in workspace is de-identified before it's used to train or refine our models. Our Privacy Policy explains how this works.

Our AI providers are barred from training on your Content.


Who can opt in

The opt-in applies to the workspace as a whole, not to individual people. The workspace's administrators give or withdraw it. If you can't change the setting, ask one of your workspace's administrators.


How to opt in or withdraw the opt-in

  1. Click your name in the bottom left of your workspace.

  2. Click Settings, then General under Workspace.

  3. Scroll to Data Usage.

  4. To opt in, select Opt in this workspace. To withdraw the opt-in, select Don't opt in.

  5. Click Save changes.

Data Usage section in HappyScribe workspace settings, showing the Opt in this workspace and Don't opt in options.

Administrators can also opt in or withdraw the opt-in by emailing dataprotection@happyscribe.com.


What withdrawing the opt-in does

Withdrawing applies from that point on. It can't undo the use of Content that was already de-identified before the opt-in was withdrawn.


Questions?

If something isn't covered here, don't hesitate to reach out by email at support@happyscribe.com

Did this answer your question?