Sony sues AI music generator Udio over 30,000 songs, seeking up to $4.5 billion in damages

Sony Lawsuit Escalates Against AI Music Generator Udio

Sony Music Entertainment has launched a second major legal action against Udio AI, filing a lawsuit in New York court that accuses the AI music generator of copyright infringement involving 30,117 unlicensed recordings1

1

. The complaint includes iconic tracks spanning decades, from Elvis Presley's "Hound Dog" to Beyoncé's "Say My Name" and Harry Styles' "As It Was," as well as works by Alicia Keys, Dolly Parton, Bob Dylan, Britney Spears, and Michael Jackson2

2

. Sony is demanding a jury trial and seeking statutory damages of up to $150,000 per infringed work, which could potentially total $4.5 billion3

3

.

From Original Complaint to Expanded Legal Battle

The new lawsuit stems from Sony's attempt to expand its original 2024 case, which it filed alongside Universal Music Group and Warner Records against both Udio and another AI music-maker, Suno1

1

. That initial complaint covered 333 works. After gaining access to Udio's training data through discovery, Sony claims it identified additional sound recordings using an "audio fingerprinting" technique1

1

. When Sony attempted to add the more than 30,000 newly identified songs to the original lawsuit, a judge rejected the motion on June 29, citing proximity to the end of the discovery phase3

3

. However, the court acknowledged that "plaintiffs have the right to seek to stop infringement of, and recover damages for, all copyrighted works," effectively clearing the path for this separate legal action2

2

.

Allegations of YouTube Data Harvesting to Train AI Models

In the new complaint, Sony Music Entertainment specifically accuses Udio of obtaining audio recordings by ripping them from YouTube to train AI models. Sony claims that Udio "admitted that its generative AI models were 'constructed by showing the program a vast amount of different kinds of sound recordings,'" including content sourced from YouTube1

1

. Udio has previously acknowledged using audio data from YouTube for training purposes but maintains that this practice constitutes fair use2

2

. The company's position on fair use stands in direct opposition to Sony's argument that such data harvesting represents willful copyright infringement on a massive scale.

Sony Stands Alone as Other Labels Strike Licensing Deals

What makes this Sony lawsuit particularly significant is that Sony now stands as the sole major label pursuing legal action against Udio. Universal Music Group and Warner Records, which were co-plaintiffs in the original 2024 lawsuit, have since struck licensing deals with the AI company1

1

. When Warner signed its agreement with Udio, the label characterized the partnership as a potential new income stream for artists who opt in2

2

. Sony, however, has taken a harder line, writing in its lawsuit that Udio's "belated embrace of licensing only underscores the unlawfulness of its decision to copy [Sony's and other labels'] copyrighted sound recordings, without a license, in the first place". This divergence in strategy highlights the music industry's split approach to AI-generated content, with some embracing collaboration while others pursue aggressive intellectual property protection.

Implications for AI Training and Copyright Law

The outcome of this case could establish a major precedent for how AI companies are permitted to use copyrighted material when they train AI models3

3

. Sony notes in its filing that the 30,117 songs represent "only a small portion of Plaintiffs' works that Udio infringed," suggesting the company could expand the lawsuit further1

1

. The company is asking the court to block Udio from infringing its works and to award damages that could reshape the economics of AI development in the creative sector. As the music industry increasingly embraces partnerships with AI companies, Sony's aggressive stance raises questions about whether licensing deals or litigation will ultimately define the relationship between traditional content owners and AI innovation. The case will likely influence how other AI companies approach training data acquisition and whether they prioritize securing licenses before deploying their models.