Independent musicians are suing Google over Lyria 3, and Google's motion to dismiss reveals more about its strategy than its answer.
A group of independent musicians filed suit alleging Google trained its Lyria 3 music AI on songs they uploaded to YouTube without authorization. Google responded with a motion to dismiss, arguing the case rests on an "unsupported hypothesis" that their specific works were used. Its defense also leans on the broad license that YouTube's terms of service grant Google over uploaded content. Google has not denied using YouTube videos for AI training broadly; it is contesting only whether these plaintiffs can prove their specific tracks were included.
The terms-of-service angle is the sharpest part of Google's defense. By uploading to YouTube, creators may have already consented to exactly this use without knowing it, which makes the lawsuit harder to sustain than it looks. This case joins a growing stack of AI training suits where the central obstacle is individual traceability: alleging a model trained on a massive dataset is easy; proving your specific file was in it is not. If Google's motion succeeds, it skips discovery entirely, the phase where it would have to document what actually fed Lyria.
The musicians' best outcome might not be winning outright. It might be surviving long enough to force Google to answer the question it is currently ducking.