YouTube AI Training Claim Sparks Indie Backlash
Alphabet’s assertion that its terms of service permit using music uploaded to YouTube for training its Lyria 3 AI model has drawn sharp criticism from independent artist advocates. The parent company of Google and YouTube stated earlier this month that it has the right, though it did not confirm it has exercised it. The policy would apply only to independent musicians and producers who upload directly, not to major rightsholders with separate contracts.
Indie artists and rightsholders often lack the leverage or budget to negotiate alternative terms, leaving them with the choice of accepting the terms or removing their music from the world’s largest streaming platform. In response, A2IM (American Association of Independent Music) CEO Ian Harrison, Music Artists Coalition executive director Ron Gubitz, and AI-campaigner Ed Newton-Rex voiced strong objections.
“It’s outrageous,” Newton-Rex said. “And I suspect if people aren’t angry about it, they don’t know about it.”
Facebook Tests AI-Powered Creator Companion App
Facebook is reimagining its Facebook Creator Studio as a standalone “AI-powered companion app to help you grow on Facebook.” The new app, currently in testing with selected creators, will provide:
- creator-specific tips from an AI assistant
- an AI-powered comment-reply drafting tool
- highlighted priority comments to reply to
It also aims to simplify data analysis: creators can ask natural language questions and receive answers and recommendations instead of navigating dashboards manually.
Eros Innovation Launches AI-Native Label with Human-Composed Music
India’s Eros Innovation has introduced Eros Music Worlds, a label powered by what it calls a “Large Cultural Model.” The label features seven AI-native artists who perform music composed by humans. The project’s output will span singles, videos, microdramas, and live experiences. Eros Innovation has also partnered with the family of late playback singer Mohammed Rafi to create new recordings and a live concert experience.
Modulate Debuts Multi-Model AI Music Detection Tool
Modulate has launched a detection model that it says “helps determine whether music contains AI-generated vocals, instrumentals, or other musical elements.” The system runs three independent models: one for identifying AI-generated vocals, one for speech, and one for instrumentals. It then combines these indicators for a more reliable assessment. The tool is aimed at distribution platforms, music streaming services, and rights organisations that are either restricting AI-generated music or allowing it onto their servers, depending on internal policies.
