Skip to content
IFAI

Industry

Musicians sue UMG and Warner over AI training deals

Published Reach 31

The American Federation of Musicians just filed a lawsuit against Universal Music Group and Warner Music, accusing the record giants of licensing members' recordings to AI music startups Suno and Udio without paying or crediting session players. The union argues these deals trigger a "new use" clause under collective bargaining agreements, demanding damages and full disclosure of the training data used.

Musicians sue UMG and Warner over AI training deals
Image: techcrunch.com

Record labels spent years warning that AI was stealing musicians' work, then signed licensing deals to monetize it. Now, the musicians who actually played those notes are dragging the labels to court.

The American Federation of Musicians (AFM) filed a federal lawsuit in the US District Court for the Southern District of New York against Universal Music Group (UMG) and Warner Music Group (WMG). The union alleges that the two music giants licensed recordings featuring its members to AI music platforms Suno and Udio without offering a dime in compensation or giving any credit.

Represented by attorney Eyad Asad, the union is seeking unspecified monetary damages and an order forcing the labels to disclose every single recording handed over to train the AI models.

The "new use" clause comes home to roost

To understand the union's fury, you have to look at how the labels handled Suno and Udio in the first place.

Back in 2024, the RIAA coordinated a massive copyright lawsuit alongside UMG, WMG, and Sony Music, accusing the AI generators of mass infringement. But by late 2025, the courtroom hostility turned into lucrative business deals.

UMG settled with Udio in late October 2025 for a payout and forward-looking licenses for a platform launching in 2026. WMG settled with Udio in mid-November, and just days later became the first major to strike a settlement with Suno, even packaging concert-discovery platform Songkick into the transaction.

The AFM argues these deals triggered the "new use" provision in its collective bargaining agreement. That clause requires major labels to compensate session musicians whenever their recorded performances are adapted into a new commercial application.

Instead, the AFM claims, the labels kept the settlement checks and future revenue streams for themselves while stonewalling the union on which catalogs were fed into the training sets.

In response, a UMG spokesperson noted that the union chose litigation while collective bargaining negotiations were actively underway, adding that the label intends to resolve the issues through continued talks.

Why it matters

This lawsuit exposes the dirty secret behind big media's AI licensing spree: signing deals with AI labs doesn't mean the money trickles down to the people whose labor built the data.

For session players—the unsung guitarists, brass sections, and drummers who make tracks sing—the stakes are existential. Without direct royalties or master ownership, collective bargaining protections are their only safety net. If ingestion into training sets isn't legally recognized as a commercial new use, human players are effectively forced to train their own automated replacements for free.

The labels played a clever hand at first. They weaponized copyright to squeeze settlements out of startups, then pivoted to ongoing licensing fees. But that playbook created a massive transparency blind spot. The Music Artists Coalition questioned artist compensation the day after UMG's Udio deal, and now the AFM is turning that skepticism into legal action.

We think this case will set a critical precedent across the entire creative sector. If the court rules that model training constitutes a new commercial use under labor agreements, record companies won't be able to pocket AI licensing windfalls alone. AI music startups might have secured catalog deals, but the real cost of their training data is likely about to go up.

Sources (1)

  1. [1]
    musicbusinessworldwide.com(opens in a new tab)Originalmusicbusinessworldwide.com

Best of AI, daily