Oleksandr Kuzmenko, EditorPublished 4 min readAI-assisted · editor-reviewed
TL;DR
- Voice data opt-in is strictly independent of Claude Code and chat training settings.
- The toggle defaults to disabled in user privacy settings.
- Voice chat recordings can be purged from Anthropic settings at any time.
Key facts
- Default Status
- Disabled
- Scope Separation
- Independent from Claude Code & chat data
Granular Training Controls
Anthropic has updated Claude's privacy structure by introducing an isolated toggle specifically for voice conversations. The interface prompt asks: *'Allow us to use your voice data to improve our AI models.'*
Crucially, this toggle is decoupled from textual chats and agent sessions:
- You can exclude your regular chat and
Claude Codeworkspace interactions from training while opting into voice testing, or vice versa. - The toggle is disabled by default in current rollouts.
Verifying Your Workspace Settings
To ensure proprietary audio and development discussions remain private, navigate to:
1. Settings > Privacy 2. Locate Allow us to use your voice data (labeled: *'For approved sessions, allow the use of your audio recordings and voice chat data to improve Anthropic AI models'*). 3. Confirm the option is toggled off, or delete prior voice data records using the in-settings deletion control.
What to do today
- Open Claude Settings > Privacy and review the state of 'Allow us to use your voice data'.
- Verify your Claude Code and workspace chat training preferences remain configured according to your team's compliance requirements.
Step names are taken from What to do today.
Step names are taken from What to do today. Open Claude Settings > Privacy and review the state of 'Allow us to use your voice data'.. Verify your Claude Code and workspace chat training preferences remain configured according to your team's compliance requirements.
- Open Claude Settings > Privacy and review the state of 'Allow us to use your voice data'.
- Verify your Claude Code and workspace chat training preferences remain configured according to your team's compliance requirements.
What the community says
I can't imagine most people will opt in to that. I hope.
