Photo credits: Phát Trương
AI music firm Suno has requested that a Massachusetts federal court reject a bid by Universal Music Group and Sony Music to add over 61,000 recordings to their copyright lawsuit. In a June 4 filing, Suno argued the labels had been delayed too long to expand the case, which originally involved 560 works.
UMG and Sony seek to add 61,026 recordings to the suit after identifying their music in training data via Audible Magic’s fingerprinting technology. Suno opposes this, arguing the expansion would necessitate further discovery, raise costs, and stall its fair-use defence.
Concurrently, a conflict persists regarding the confidentiality of Suno’s training dataset size. Suno aims to seal these details to prevent competitive harm, though it has admitted its model was likely trained on tens of millions of recordings, including those owned by the plaintiffs.
Initiated in 2024 by the RIAA on behalf of major labels, this pivotal lawsuit explores copyright law’s application to generative AI training. While Warner Music Group has settled, the remaining plaintiffs, UMG and Sony continue the case.
The debate over AI training data disclosure intensified after a June 3 ruling in the Udio copyright case. U.S. District Judge Alvin Hellerstein vacated an order that had sealed parts of Sony Music’s lawsuit, questioning whether Udio must reveal confidential details like its training dataset size, as reported by Music Business Worldwide. Sony claims this data is vital to its allegations that 30,000 recordings were ingested, while Udio argues disclosure would aid competitors.
