ElevenLabs Signs Partnership with UK AI Security Institute (AISI)

ElevenLabs has signed a partnership with the UK government lasting 3 years. Through this collaboration, they will research the impact of AI voice technology on society and security, specifically investigating how well people can identify when they are talking to an AI agent, and how their behavior changes depending on whether they are aware of this.

For the research, ElevenLabs will provide the UK government with access to its frontier voice models. The research team plans to focus on the ability to distinguish between humans and AI during natural conversations, and how willingness to disclose information changes when someone knows the other party is an AI.

ElevenLabs operates the following safeguards to prevent misuse of the technology:

• Blocking voice cloning of politicians and celebrities
• Requiring identity verification for voice cloning
• Providing an AI Audio Classifier to determine whether audio is AI-generated

This partnership is an extension of the recently launched ElevenLabs for Government. Already, several public institutions, including the Ukrainian government, are using this technology to provide essential services to citizens or handle large-scale call center operations.