xAI Launches Custom Voices Feature for Voice Cloning from One Minute of Speech

xAI has released a Custom Voices feature that enables developers to clone voices for AI applications using just one minute of speech input. The feature builds on xAI's recently launched Grok Speech-to-Text and Text-to-Speech APIs, expanding the company's voice technology offerings. The low barrier of only one minute of audio required for voice cloning makes the technology highly accessible to developers building AI applications. This release positions xAI in the growing market for voice synthesis and cloning tools, though the ease of cloning voices raises potential concerns about misuse for impersonation or deepfakes.

xAI's Custom Voices Feature and Technical Capabilities

• ▪xAI launched a Custom Voices feature that allows developers to clone their own voices for AI applications.
• ▪xAI's Custom Voices feature can create a usable voice clone from one minute of speech.

Integration with Existing xAI Speech APIs

• ▪xAI's Custom Voices feature builds on the recently launched Grok Speech-to-Text API.

Perspective of AI safety researchers and ethicists

• ▪Voice cloning technology like xAI's Custom Voices feature could enable creation of convincing audio deepfakes for fraud or misinformation.
• ▪xAI's Custom Voices feature requiring only one minute of speech raises concerns about unauthorized voice cloning and impersonation risks.

Perspective of Developers and AI application builders

• ▪The one-minute speech requirement of xAI's Custom Voices feature lowers the barrier to entry for implementing voice cloning in developer projects.
• ▪xAI's Custom Voices feature provides developers with an efficient way to create personalized voice interfaces for AI applications.

1 source