Anthropic has started asking Claude users to voluntarily share their voice conversations to help train and improve its AI models, according to BleepingComputer. A prompt appears when people use Claude’s voice features and reads “Allow us to use your voice data to improve our AI models,” explaining that audio recordings and voice chat data help the models understand and respond to speech, and that users can turn the setting off or delete the data at any time in settings.
The choice is optional: users can pick “Allow” or “Not now,” and there is a dedicated toggle under Settings > Privacy labelled “Allow us to use your voice data,” which applies to “approved sessions.” In the interface BleepingComputer saw, the setting is disabled by default, so users are not automatically opted in. Voice training is separate from Anthropic’s existing option to use chats and Claude Code sessions for training, which means someone could allow one and not the other. Voice recordings can carry more personal information than typed text, such as a person’s voice itself and anything said aloud nearby, so it is worth knowing which way the toggle is set.
How to check if you’re affected
Affected products are Claude apps that offer voice features. Open Settings > Privacy and look for the voice data option (“Allow us to use your voice data”), and confirm it is set the way you want. If you tapped “Allow” on the prompt and have changed your mind, switch the toggle off and use the delete option Anthropic describes in settings to remove existing voice data. Check the chat and Claude Code training options in the same Privacy screen separately, since they are independent controls.