ElevenLabs announced on the 23rd new features for its AI music generation service "ElevenMusic," including "Reference" and "Vocal." Reference generates a new track matching the mood and style of the AI model "Music v2" when a user uploads an audio clip 10 seconds to 5 minutes long. Vocal creates a fully original song sung in a given voice using the user's voice or a library voice, and can learn a specific voice's tone and singing style to maintain the same vocal across multiple songs. The "Style" feature now also supports "Music version 2." An ElevenLabs official stated that all voices and tracks uploaded are designed to be usable only after passing a strict copyright check before the generation stage, and generation using another person's copyrighted work is blocked.