Anthropic Expands Claude Voice Capabilities with Data Training Opt-In

Anthropic is inviting Claude users to share voice conversations to help train its AI models, unveiling a new voice data consent option as the company expands voice mode across more capable language models. The company began rolling out the opt-in setting on October 4, 2026, asking users to permit the use of voice recordings and audio chat data to improve its AI systems. The feature operates independently of Anthropic's existing text-training consent, allowing users to authorize voice training while keeping text conversations excluded, or vice versa. The voice data setting defaults to off, remaining unavailable until users explicitly enable it in Claude's Privacy settings.