Overview
- Google filed a 41-page motion to dismiss on June 10, 2026, saying YouTube’s terms of service give it a worldwide, royalty-free right to use and process uploaded content and that the plaintiffs cannot prove Google trained on their specific songs.
- A proposed class action filed in March by independent artists alleges Google used YouTube uploads without consent to train its Lyria 3 music model and reported that roughly 44 million audio clips were scraped, a figure presented as an allegation in the complaint.
- Lyria 3, released inside Google’s Gemini app, can generate short music tracks with vocals and lyrics and the model’s commercial capability is central to the bands’ claim that the AI competes with their work.
- Plaintiffs say Google had ‘structural leverage’ through ownership of YouTube and Content ID and could have licensed music instead of copying, while Google argues the platform’s licensing language covers the disputed use.
- If the court rejects Google’s terms-of-service defense the decision could force mass licensing or new data strategies for companies that train models on user uploads and will determine the scope of discovery and potential damages in the case.