Sony Music Files Another Lawsuit Against Udio, Alleges AI Music Generator Copied 30,000 Songs to Train Models
Sony Music Entertainment has filed a fresh lawsuit against AI music generator Udio, accusing it of unlawfully copying more than 30,000 songs by artists including Beyoncé, Harry Styles and Elvis Presley to train its models. The new complaint, lodged on 20 July 2026 in New York, follows a judge's refusal to let Sony add the newly discovered tracks to its original 2024 suit, prompting the label to launch a separate case instead. It forms part of a wider legal battle between major record labels and AI music firms over how copyrighted recordings can be used to train generative models.
Sony's original action against Udio and rival Suno was brought jointly with Universal Music Group and Warner Music Group in June 2024, but Universal and Warner have since settled and struck licensing deals with Udio, which is preparing an industry-backed model for release later this year. Sony alleges discovery revealed 30,117 of its tracks in Udio's training data, taken from YouTube without permission; Udio has argued such use constitutes fair use. Sony is seeking at least $150,000 per infringed work and a court order blocking Udio from using its catalogue to train future models.
- Sony sues Udio again, alleging 30,000+ songs used without permission
- Follows judge's rejection of adding tracks to original 2024 lawsuit
- Sony seeks $150,000 per song and a training injunction
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Advocates for Sony's position argue that copyright exists precisely to ensure creators and rights-holders are compensated when their work is used commercially, and that scraping tens of thousands of copyrighted recordings from YouTube without a licence, then using them to build a product that competes with the original artists, is straightforwardly unauthorised commercial exploitation rather than fair use. They point to Universal and Warner's decision to strike licensing deals as evidence that a workable, consent-based model exists, and argue that allowing AI firms to bypass it would gut incentives for musicians and labels to keep investing in new recordings.
The case against
Advocates for Udio's position argue that training a generative model on existing works to learn patterns and produce new, non-identical output is a transformative use analogous to how human musicians learn by listening to and absorbing the music around them, and that fair use doctrine has historically protected exactly this kind of transformative, non-substitutive activity. They contend that permitting rights-holders to block model training entirely, rather than seeking proportionate licensing remedies, would hand a small number of major labels an effective veto over an emerging technology and stifle innovation and competition in AI-generated music.
AI Art Business Celebrity Companies Culture Entertainment Fashion Music Technology